About DailyArXiv
DailyArXiv is an MCP server in the Documentation category: daily ArXiv Papers.
Install
From source
git clone https://github.com/zezhishao/DailyArXivThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
✦
Ask AI about DailyArXiv
Powered by Claude · Grounded in docs
I know everything about DailyArXiv. Ask me about installation, configuration, usage, or troubleshooting.
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
Daily Papers
The project automatically fetches the latest papers from arXiv based on keywords.
The subheadings in the README file represent the search keywords.
Only the most recent articles for each keyword are retained, up to a maximum of 100 papers.
You can click the 'Watch' button to receive daily email notifications.
Last update: 2026-05-07
Time Series
| Title | Date | Abstract | Comment |
|---|---|---|---|
| A Domain Incremental Continual Learning Benchmark for ICU Time Series Model Transportability | 2026-05-05 | ShowIn recent years, machine learning has made significant progress in clinical outcome prediction, demonstrating increasingly accurate results. However, the substantial resources required for hospitals to train these models, such as data collection, labeling, and computational power, limit the feasibility for smaller hospitals to develop their own models. An alternative approach involves transferring a machine learning model trained by a large hospital to smaller hospitals, allowing them to fine-tune the model on their specific patient data. However, these models are often trained and validated on data from a single hospital, raising concerns about their generalizability to new data. Our research shows that there are notable differences in measurement distributions and frequencies across various regions in the United States. To address this, we propose a benchmark that tests a machine learning model's ability to transfer from a source domain to different regions across the country. This benchmark assesses a model's capacity to learn meaningful information about each new domain while retaining key features from the original domain. Using this benchmark, we frame the transfer of a machine learning model from one region to another as a domain incremental learning problem. While the task of patient outcome prediction remains the same, the input data distribution varies, necessitating a model that can effectively manage these shifts. We evaluate two popular domain incremental learning methods: data replay, which stores examples from previous data sources for fine-tuning on the current source, and Elastic Weight Consolidation (EWC), a model parameter regularization method that maintains features important for both data sources. | |
| Conditional independence testing with a single realization of a multivariate nonstationary nonlinear time series | 2026-05-05 | ShowIdentifying relationships among stochastic processes is a core objective in many fields, such as economics. While the standard toolkit for multivariate time series analysis has many advantages, it can be difficult to capture nonlinear dynamics using linear vector autoregressive models. This difficulty has motivated the development of methods for causal discovery and variable selection for nonlinear time series, which routinely employ tests for conditional independence. In this paper, we introduce the first framework for conditional independence testing that works with a single realization of a nonstationary nonlinear process. We also show how our framework can be used to test for independence. The key technical ingredients of our framework are time-varying nonlinear regression, estimation of local long-run covariance matrices of products of error processes, and a distribution-uniform strong Gaussian approximation. | |
| Training-Free Probabilistic Time-Series Forecasting with Conformal Seasonal Pools | 2026-05-05 | ShowWe propose Conformal Seasonal Pools (CSP), a training-free probabilistic time-series forecaster that mixes same-season empirical draws with signed residual draws around a seasonal naive forecast. In an audited rolling-origin benchmark on the six time-series datasets where DeepNPTS was originally evaluated (electricity, exchange_rate, solar_energy, taxi, traffic, wikipedia), CSP-Adaptive significantly outperforms DeepNPTS on every metric we report -- CRPS (per-window paired Wilcoxon $p \approx 4 \times 10^{-10}$), normalized mean quantile loss ($p \approx 7 \times 10^{-10}$), and empirical 95% coverage ($p \approx 8 \times 10^{-45}$, mean 0.89 vs 0.66) -- while running over 500x faster on CPU. Coverage is the most decision-critical of these: a 0.95 nominal interval that contains the truth in only ~66% of cases fails the basic calibration desideratum and would not survive deployment in safety- or decision-critical settings. The failure mode is also more severe than aggregate coverage suggests: in the worst 10% of windows, DeepNPTS's prediction interval covers none of the H forecast horizons -- the entire multi-step trajectory misses the truth at every step simultaneously. This poses serious risk in safety- and decision-critical applications such as healthcare, finance, energy operations, and autonomous systems, where prediction intervals that systematically miss the truth across the entire planning horizon translate directly into misclassified patients, regulatory capital failures, grid imbalances, and safety-case violations. CSP achieves all of this with no learned parameters and no training. We argue training-free conformal samplers should be mandatory baselines when evaluating learned non-parametric forecasters. | |
| Segmenting Human-LLM Co-authored Text via Change Point Detection | 2026-05-05 | ShowThe rise of large language models (LLMs) has created an urgent need to distinguish between human-written and LLM-generated text to ensure authenticity and societal trust. Existing detectors typically provide a binary classification for an entire passage; however, this is insufficient for human--LLM co-authored text, where the objective is to localize specific segments authored by humans or LLMs. To bridge this gap, we propose algorithms to segment text into human- and LLM-authored pieces. Our key observation is that such a segmentation task is conceptually similar to classical change point detection in time-series analysis. Leveraging this analogy, we adapt change point detection to LLM-generated text detection, develop a weighted algorithm and a generalized algorithm to accommodate heterogeneous detection score variability, and establish the minimax optimality of our procedure. Empirically, we demonstrate the strong performance of our approach against a wide range of existing baselines. | |
| Long-Range Correlation in Code Commit Dynamics as a Novel Indicator of Software Product Stability: A Detrended Fluctuation Analysis Study | 2026-05-05 | ShowThis work proposes the fractal scaling exponent alpha, estimated via Detrended Fluctuation Analysis (DFA) on the unaggregated time series of lines of code added per commit event in a software repository, as a novel process-level indicator of software product stability. The proposal rests on the hypothesis that stable software products arise from development processes characterised by long-range temporal correlations in commit behaviour: each code addition is shaped not only by the immediately preceding commits but by patterns extending weeks or months into the past and anticipating work to be done in the future. This hypothesis is tested on two non-overlapping 712-day time series of lines of code added per commit event, drawn from a closed-source software organisation and labeled as stable and unstable by the lead engineer on the basis of crash-analytics data. Applied to these series, DFA yields alpha = 0.70 (n_min = 16) for the stable period and alpha = 0.57 for the unstable period, with all estimates substantially above the shuffled-surrogate baseline (alpha ~= 0.50 +/- 0.01). Results are robust to three parameterisations (n_min in {4, 16, 48}) and validated against 1,000 surrogate time series per condition. Remarkably, the unstable period generated 3.2 times more commit events than the stable period, yet exhibited lower long-range memory, demonstrating that commit volume alone does not predict stability, and that the temporal organisation of development activity is the key variable. This result can be situated in the broader literature on fractality in human creative production, discuss methodological limitations, and outline a research programme for deploying alpha as a continuous code-health indicator in version-control pipelines. | 17 pages, 5 figures |
| FinSTaR: Towards Financial Reasoning with Time Series Reasoning Models | 2026-05-05 | ShowTime series (TS) reasoning models (TSRMs) have shown promising capabilities in general domains, yet they consistently fail on financial domain, which exhibit unique characteristics. We propose a general 2x2 capability taxonomy for TSRMs by crossing 1) single-entity vs. multi-entity analysis with 2) assessment of the current state vs. prediction of future behavior. We instantiate this taxonomy in the financial domain -- where the distinction between deterministic assessment and stochastic prediction is particularly critical -- as ten financial reasoning tasks, forming the FinTSR-Bench benchmark based on S&P stocks. To this end, we propose FinSTaR (Financial Time Series Thinking and Reasoning), trained on FinTSR-Bench with distinct chain-of-thought (CoT) strategies tailored to each category. For assessment, which is deterministic (i.e., computable from observable data), we employ Compute-in-CoT, a programmatic CoT that enables models to derive answers directly from raw prices. For prediction, which is inherently stochastic (i.e., subject to unobservable factors), we adopt Scenario-Aware CoT, which generates diverse scenarios before making a judgment, mirroring how financial analysts reason under uncertainty. The proposed method achieves 78.9% average accuracy on FinTSR-Bench, substantially outperforming LLM and TSRM baselines. Furthermore, we show that the four capability categories are complementary and mutually reinforcing through joint training, and that Scenario-Aware CoT consistently improves prediction accuracy over standard CoT. Code is publicly available at: https://github.com/seunghan96/FinSTaR. | |
| Using Echo-State Networks to Reproduce Rare Events in Chaotic Systems | 2026-05-05 | ShowWe apply Echo-State Networks to predict time series and statistical properties of the competitive Lotka-Volterra model in the chaotic regime. In particular, we demonstrate that Echo-State Networks successfully learn the chaotic attractor of the competitive Lotka-Volterra model and reproduce histograms of dependent variables, including tails and rare events. We also demonstrate that the Echo-State Networks reproduce rare events in the non-equilibrium simulations of the Lotka-Volterra system. We use the Generalized Extreme Value distribution to quantify the tail behavior. | |
| Global and Local Topology-Aware Attention with Persistent Homology and Euler Biases for Time-Series Forecasting | 2026-05-04 | ShowScientific time series often encode predictive geometric structure, including connectivity, cycles, shell-like geometry, directional changes, and nonlinear neighborhoods, that standard dot-product attention does not explicitly represent. We introduce a topology-aware attention framework that adds such structure to attention logits using persistent homology (H0-H2), anchored Euler characteristic transforms, and kernel-Hilbert channels. A validation-gated local residual captures local topological signals, including a Zeng-style local H0 component, only when held-out validation data support the correction. Exact Vietoris-Rips computations and smooth topological surrogates are evaluated under a no-leakage protocol with train-only calibration, validation-only selection, and test-only reporting. We evaluate guarded topology-aware variants across three architecture families: lightweight attention/Ridge, PatchTSTForRegression, and TimeSeriesTransformerForPrediction. Experiments include synthetic benchmarks isolating higher-order topology and real datasets covering CO2, S&P 500 return-window geometry, and NASA IMS bearing degradation. The audit uses matched paired comparisons across seven dataset units, three random seeds, and three chronological splits, giving 63 paired units per architecture and 189 paired units overall. Topology-aware models show positive paired effects when geometry is predictive, with heterogeneous magnitude across datasets and architectures. Lightweight attention/Ridge improves in 46 of 63 units, with mean relative RMSE reduction of 12.5% and paired randomization p=7.2e-4; PatchTST improves in 33 units and retains the baseline in 20 units, with 23.5% reduction and p=3.5e-5; and TimeSeriesTransformer improves in 47 units, with 47.8% reduction and p<1e-4. The results support topology as a validation-selected, architecture-compatible inductive bias. | |
| TCD-Arena: Assessing Robustness of Time Series Causal Discovery Methods Against Assumption Violations | 2026-05-04 | ShowCausal Discovery (CD) is a powerful framework for scientific inquiry. Yet, its practical adoption is hindered by a reliance on strong, often unverifiable assumptions and a lack of robust performance assessment. To address these limitations and advance empirical CD evaluation, we present TCD-Arena, a modularized, highly customizable, and extendable testing kit to assess the robustness of time series CD algorithms against stepwise more severe assumption violations. For demonstration, we conduct an extensive empirical study comprising around 30 million individual CD attempts and reveal nuanced robustness profiles for 33 distinct assumption violations. Further, we investigate CD ensembles and find that they have the potential to improve general robustness, which has implications for real-world applications. With this, we strive to ultimately facilitate the development of CD methods that are reliable for a diverse range of synthetic and potentially real-world data conditions. | |
| The Bayesian Reflex: Online Learning as the Autonomic Nervous System of Modern and Future AI | 2026-05-04 | ShowThis chapter introduces the Bayesian reflex -- an analogy with the autonomic nervous system -- as a unifying framework for online learning in AI. Bayesian online algorithms automatically maintain equilibrium in dynamic environments via three mechanisms: belief maintenance through probabilistic representations, sequential updating via Bayes' theorem, and uncertainty-driven action balancing exploration and exploitation. We survey online Bayesian methods, highlighting two computational principles: the look-up table principle for sequential inference in function space, and the ellipsoidal decomposition framework for nearly exact i.i.d. sampling from arbitrary posteriors. These principles are generalized across dynamic emulation, nonparametric state-space models, circular time series, inverse regression for climate model evaluation, and deep architectures via Recursive Gaussian Processes. Decision-making is explored via Thompson sampling and restless bandits. We extend the framework to assess infinite series convergence (applied to climate dynamics and the Riemann Hypothesis), model prime number distributions leading to the discovery of 184 strong Mersenne prime candidates, detect stationarity, and characterize point processes. The Bayesian reflex provides a foundational infrastructure for adaptive AI that continuously learns in a complex world. | Feedback welcome |
| Period-conscious Time-series Reconstruction under Local Differential Privacy | 2026-05-04 | ShowPeriodic patterns are fundamental cues in multimedia signals and systems, including repetitive motion in video (e.g., gait cycles), rhythmic and pitch-related structure in audio, and recurring textures in image sequences. When such user-generated streams are collected from edge devices, local differential privacy (LDP) is appealing because it perturbs data before upload; however, the injected noise can corrupt spectral peaks and induce phase drift, making period estimation unreliable and degrading reconstruction quality. We propose \textbf{CPR} (\textit{Cycle and Phase Recovery}), a period-aware reconstruction framework for periodic time series under LDP. CPR performs multi-scale period probing and multi-consensus selection to suppress noise-induced spectral interference, then aggregates perturbed samples at matched within-cycle phase positions to stabilize phase alignment across cycles. To recover the underlying per-phase values, CPR combines EM-based denoising with kernel density estimation, improving robustness under tight privacy budgets. Experiments on two real-world periodic datasets demonstrate that CPR better preserves periodic structure and consistently achieves lower reconstruction error than representative LDP baselines, especially in the low-$ε$ regime. | |
| Beyond Sequential Prediction: Learning Financial Market Dynamics in Volatile and Non-Stationary Environments through Sentiment-Conditioned Generative Modelling | 2026-05-04 | ShowThe problem of time-series forecasting in non-stationary and complex environments is a challenging task in machine learning, especially with heterogeneous numerical and textual data present. Traditional statistical models like AutoRegressive Integrated Moving Average (ARIMA) are based on the assumptions of linearity and stationarity, whereas recurrent neural networks like Long Short-Term Memory (LSTM) models do not necessarily represent distributional properties in highly volatile settings. This paper proposes a hybrid model that combines Generative Adversarial Networks (GANs) with Natural Language Processing (NLP)-based sentiment analysis to enable sentiment-conditioned time-series prediction. The model integrates adversarial learning on numerical sequences with contextual sentiment representations derived from unstructured text, enabling them to be jointly modelled to capture temporal dynamics and exogenous information. These results demonstrate the promise of hybrid generative and language-aware methods to enhance prediction robustness in non-stationary environments. | 18 pages, 5 figures |
| MSMixer: Learned Multi-Scale Temporal Mixing with Complementary Linear Shortcut for Long-Term Time Series Forecasting | 2026-05-04 | ShowLong-term time series forecasting requires models that simultaneously capture rapid oscillations, medium-range periodicities, and slowly evolving macro-trends from a fixed look-back window. Existing lightweight MLP-based models typically operate on a single temporal resolution, limiting their ability to explicitly model patterns at multiple scales. We propose MSMixer, a channel-independent multi-scale MLP architecture that addresses this limitation through three complementary innovations: (i) three parallel scale branches at down-sample factors {1x, 4x, 16x} with independent MLP blocks, (ii) a learnable softmax gate that dynamically weighs branch outputs, and (iii) a DLinear complementary shortcut that provides full-window trend and seasonality context. MSMixer contains only 112K parameters at H=96 and runs at O(T) complexity. Evaluated on four ETT benchmarks with standard chronological splits and three random seeds, MSMixer achieves the lowest average MSE (0.357) among lightweight models, outperforming DLinear (0.386, -7.4%) and NLinear (0.365, -2.1%), winning 12 of 16 configurations. Against five Transformer-based baselines from the literature, MSMixer achieves best or second-best MSE in 9 of 16 configurations while using 5x fewer parameters than PatchTST. Ablation and sensitivity analyses confirm the complementary contributions of the multi-scale branches and the DLinear shortcut. | 21 pa...21 pages, 5 figures, 8 tables. Submitted to International Journal of Machine Learning and Cybernetics (Springer) |
| Online Generalised Predictive Coding | 2026-05-04 | ShowThis paper introduces an extension of generalised filtering for online applications. Generalised filtering refers to data assimilation schemes that jointly infer latent states, learn unknown model parameters, and estimate uncertainty in an integrated framework -- e.g., estimate state and observation noise -- at the same time (i.e., triple estimation). This framework appears across disciplines under different names, including variational Kalman-Bucy filtering in engineering, generalised predictive coding in neuroscience, and Dynamic Expectation Maximisation (DEM) in time-series analysis. Here, we specialise DEM for ``online'' data assimilation, through a separation of temporal scales. We describe the variational principles and procedures that allow one to assimilate data in a way that allows for a slow updating of parameters and precisions, which contextualise fast Bayesian belief updating about the dynamic hidden states. Using numerical studies, we demonstrate the validity of online DEM (ODEM) using a non-linear -- and potentially chaotic -- generative model, to show that the ODEM scheme can track the latent states of the generative process, even when its functional form differs fundamentally from the dynamics of the generative model. Framed from a neuro-mimetic predictive coding perspective, ODEM offers a biologically inspired solution to online inference, learning, and uncertainty estimation in dynamic environments. | 45 pages, 17 Figures |
| EstemPMM: Polynomial Maximization Method for Non-Gaussian Regression and Time Series in R | 2026-05-04 | ShowWe describe the R package EstemPMM, which implements the Polynomial Maximization Method (PMM) for parameter estimation under non-Gaussian errors. PMM exploits higher-order cumulants of the error distribution -- specifically the third standardized moment gamma_3 and fourth standardized moment gamma_4 -- to construct estimators that outperform ordinary least squares (OLS) whenever the errors are asymmetric or leptokurtic. The package provides a unified interface for linear regression (lm_pmm2, lm_pmm3), autoregressive and moving-average time-series models (ar_pmm2, ma_pmm2, arma_pmm2, arima_pmm2, and seasonal variants), a data-driven dispatch function (pmm_dispatch) that automatically selects OLS, PMM2, or PMM3 based on the sample skewness and excess kurtosis, and Monte Carlo comparison utilities. The implementation uses R's S4 class system and follows standard generic interfaces (coef, fitted, residuals, predict, summary, AIC, logLik, vcov, confint). Asymptotic efficiency is characterised by Kunchenko-style coefficients g_2, g_3 in [0,1], defined as the ratios of the asymptotic variance of the PMM2 and PMM3 estimators to that of OLS. Monte Carlo experiments confirm the theoretical values and a WTI crude-oil case study illustrates the dispatcher and parameter-precision benefits of PMM2 on real heavy-tailed data. EstemPMM version 0.3.2 is available from CRAN at https://CRAN.R-project.org/package=EstemPMM under the GPL-3 licence. | R pac...R package, 9 sections, 10 figures, 9 tables. EstemPMM version 0.3.2 on CRAN |
| TokenTiming: A Dynamic Alignment Method for Universal Speculative Decoding Model Pairs | 2026-05-04 | ShowAccelerating the inference of large language models (LLMs) has been a critical challenge in generative AI. Speculative decoding (SD) substantially improves LLM inference efficiency. However, its utility is limited by a fundamental constraint: the draft and target models must share the same vocabulary, thus limiting the herd of available draft models and often necessitating the training of a new model from scratch. Inspired by Dynamic Time Warping (DTW), a classic algorithm for aligning time series, we propose the algorithm TokenTiming for universal speculative decoding. It operates by re-encoding the draft token sequence to get a new target token sequence, and then uses DTW to build a mapping to transfer the probability distributions for speculative sampling. Benefiting from this, our method accommodates mismatched vocabularies and works with any off-the-shelf models without retraining and modification. We conduct comprehensive experiments on various tasks, demonstrating 1.57x speedup. This work enables a universal approach for draft model selection, making SD a more versatile and practical tool for LLM acceleration. | Accepted by ACL 2026 |
| Learning Temporal Patterns in Financial Time Series: A Comparative Study of Quantum LSTM and Quantum Reservoir Computing | 2026-05-04 | ShowThis study explores quantum and classical hybrid architectures for financial time-series fore casting, focusing on Quantum Long Short-Term Memory (QLSTM) networks and Quantum Reservoir Computing (QRC), using univariate and multivariate lag structures on real financial data. We assess how lag embeddings affect predictive accuracy and robustness. Data are en coded into quantum states via amplitude encoding, enabling efficient representation of normalized lagged observations under realistic qubit constraints. The recurrent dynamics of QLSTM and the reservoir of QRC are implemented as parameterized quantum circuits, while classical optimizers train the readout and, where applicable, variational circuit parameters. We benchmark quantum models against classical LSTM and reservoir computing using common error like metrics. Our results show that, with suitable lag selection and amplitude encoding, quantum-enhanced archi tectures match classical baselines in univariate settings and can modestly outperform them in multivariate regimes with correlated inputs, where expressive encodings are most beneficial. | |
| Prior elicitation for Bayesian estimation of single-subject connectivity networks | 2026-05-04 | ShowInference of brain functional connectivity networks from resting-state fMRI data is a key focus in neuroimaging. This paper introduces new Bayesian approaches for inferring a functional connectivity graph from multivariate resting-state fMRI time series of a single subject. Our methods rely on novel Bayesian priors on correlation matrices and a dedicated prior elicitation framework, which translates prior beliefs about the expected level and variability of correlations into interpretable hyperparameter choices, enabling the construction of expert-informed priors. When combined with a Gaussian likelihood, these priors also exhibit computational advantages. Compared to most existing methods for this problem that estimate constant weights, our model provides distributional weights defined by the posterior distributions for the connectivity graph, yielding more robust point estimates through the regularizing effect of expert-informed priors, evaluating uncertainty, and enabling a range of post-inference analyses. In particular, we derive a procedure for identifying significant connectivities based on posterior distributions of weights and credible sets. To the best of our knowledge, only one existing Bayesian functional connectivity model is applicable to single-subject resting-state fMRI data, making our approach a valuable addition to the field and demonstrating superior performance in our experiments. | |
| Model checking with temporal graphs and their derivative | 2026-05-04 | ShowTemporal graphs are graphs where the presence or properties of their vertices and edges change over time. When time is discrete, a temporal graph can be defined as a sequence of static graphs over a discrete time span, called lifetime, or as a single graph where each edge is associated with a specific set of time instants where the edge is alive. For static graphs, Courcelle's Theorem asserts that any graph problem expressible in monadic second-order logic can be solved in linear time on graphs of bounded tree-width. We propose the first adaptation of Courcelle's Theorem for monadic second-order logic on temporal graphs that does not explicitly rely on a parameter proportional to the lifetime, or defined as the maximum number of time-edges incident with any vertex which in the worst case is higher than the lifetime. We then introduce the notion of derivative over a sliding time window of a chosen size, and define the tree-width and twin-width of the temporal graph's derivative. We exemplify its usefulness with meta-theorems with respect to a temporal variant of first-order logic. The resulting logic expresses a wide range of temporal graph problems including a version of temporal cliques, an important notion when querying time series databases for community structures. | |
| CastFlow: Learning Role-Specialized Agentic Workflows for Time Series Forecasting | 2026-05-04 | ShowRecently, large language models (LLMs) have shown great promise in time series forecasting. However, most existing LLM-based forecasting methods still follow a static generative paradigm that directly maps historical observations to future values in a single pass. Under this paradigm, forecasting is constrained by limited temporal pattern extraction, single-round acquisition of contextual features, one-shot forecast generation, and lack of support from ensemble forecasts. To address these limitations, in this work, we propose CastFlow, a dynamic agentic forecasting framework that enables multi-view temporal pattern extraction, multi-round contextual features acquisition, iterative forecast refinement, and forecasting with ensemble forecasts. First, CastFlow organizes the forecasting process into planning, action, forecasting, and reflection, establishing an agentic workflow. Second, this workflow is supported by a memory module that retrieves prior experience and a multi-view toolkit that constructs diagnostic evidence and provides a reliable ensemble forecast baseline. Third, CastFlow adopts a role-specialized design that combines general-purpose reasoning with specialized numerical forecasting. Under this design, a frozen LLM preserves general-purpose reasoning, while a fine-tuned domain-specific LLM performs evidence-guided numerical forecasting based on the ensemble forecast baseline, rather than from scratch. To optimize a fine-tuned domain-specific LLM, we further develop a two-stage workflow-oriented training that combines supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR). To evaluate the effectiveness of CastFlow, we conduct extensive experiments on diverse datasets and show that it achieves superior overall results against strong baselines. We hope that this work can serve as a step toward more adaptive and accurate time series forecasting. | |
| Exploring Accuracy Law for Deep Time Series Forecasters: An Empirical Study | 2026-05-04 | ShowDeep time series forecasting has emerged as a rapidly growing field in recent years. Despite the exponential growth of community interests, progress on standard benchmarks is often limited to marginal improvements. A common consensus of the community is that time series forecasting inherently faces a non-zero error lower bound due to its partially observable and uncertain nature. However, a fundamental question arises: how to estimate the performance upper bound of deep time series forecasters? We delve into univariate time series forecasting, a prevalent forecasting paradigm spanning traditional statistical models to advanced time series foundation models. Going beyond classical series-wise predictability metrics, we realize that the forecasting performance is highly related to window-wise properties due to the sequence-to-sequence forecasting paradigm of deep time series models and introduce a quantitative measurement of window-wise pattern complexity. Through rigorous statistical analyses over more than 4700 newly trained deep forecasting models, we discover a consistent empirical relationship between the minimum attainable forecasting error of deep models and the complexity of window-wise series patterns, which is termed the accuracy law. We further demonstrate that this empirical finding successfully guides us to identify saturated tasks from widely used benchmarks and derive an effective training strategy for time series foundation models, offering valuable insights for future research. | |
| Large-Scale Asset Selection via Metric Dependence with Enriched High Frequency Information | 2026-05-04 | ShowLarge scale portfolio choice is highly sensitive to estimation error, making the preliminary asset selection essential in empirical implementation. Existing selection rules typically rely on scalar returns or low dimensional high frequency summaries, and thus discard intraday risk dynamics that may be relevant for risk adjusted allocation. We propose Metric Dependence Screening (MDS), an asset selection procedure that incorporates high frequency information as object valued data. Each asset day observation is represented as a point-curve object combining daily return with an intraday risk state curve, equipped with a weighted product metric that preserves both reward information and within day risk dynamics. MDS ranks assets by a Fréchet variation based dependence score, measuring how much a risk adjusted target explains the metric dispersion of the asset representations. This yields a simple two stage portfolio procedure: MDS first reduces the investable universe, and standard mean-variance or minimum variance allocation is then applied. We develop a target slicing estimator and establish concentration, sure selection, and rank consistency guarantees under $α$-mixing time series dependence and ultrahigh dimensionality. Simulations show that MDS performs well across both Euclidean and non-Euclidean settings. Using 5 minute data for 2,938 Chinese A-share stocks from July 2023 to December 2025, we demonstrate that MDS improves out of sample portfolio performance over return based and scalar dependence based benchmarks, highlighting the value of preserving intraday risk dynamics. | |
| Deep Time Series Models: A Comprehensive Survey and Benchmark | 2026-05-04 | ShowTime series, characterized by a sequence of data points organized in a discrete-time order, are ubiquitous in real-world scenarios. Unlike other data modalities, time series present unique challenges in learning and modeling due to their intricate and dynamic nature, including the entanglement of nonlinear patterns and time-variant trends. Recent years have witnessed remarkable breakthroughs in time series analysis, with techniques shifting from traditional statistical methods to contemporary deep learning models. In this paper, we delve into the design of deep time series models across various analysis tasks and review the existing literature from two perspectives: basic modules and model architectures. Further, we develop and release Time Series Library (TSLib) as a fair benchmark of deep time series models for diverse analysis tasks. TSLib implements 41 prominent models, including both small- and large-scale time series models, covers 30 datasets from different domains, and supports 5 prevalent analysis tasks. Based on TSLib, we evaluate 16 popular deep time series models and 6 advanced time series foundation models. Empirical findings indicate that models with specific structures are apt only at distinct analytical tasks, providing insights for research and adoption of deep time series models. Code and datasets are available at https://github.com/thuml/Time-Series-Library. | |
| HELIX: Hybrid Encoding with Learnable Identity and Cross-dimensional Synthesis for Time Series Imputation | 2026-05-04 | ShowTime series imputation benefits from leveraging cross-feature correlations, yet existing attention-based methods re-discover feature relationships at each layer, lacking persistent anchors to maintain consistent representations. To address this, we propose HELIX, which assigns each feature a learnable feature identity, a persistent embedding that captures intrinsic semantic properties throughout the network. Unlike graph-based methods that rely on predefined topology and assume homogeneous spatial relationships, HELIX learns arbitrary feature dependencies end-to-end from temporal co-variation, naturally handling datasets where features mix spatial locations with semantic variables. Integrated with hybrid temporal-feature attention, HELIX achieves the state-of-the-art performance, surpassing all 16 baselines on 5 public datasets across 21 experimental settings in our evaluation. Furthermore, our mechanistic analysis reveals that HELIX aligns learned feature identities and dependencies with latent physical and semantic structure progressively across layers, demonstrating that it more effectively translates cross-feature structure into imputation accuracy. | Accep...Accepted at ICML 2026 (spotlight paper) |
| From 'Here' to 'There': Exploring Proximity Semantics in Multimodal Data Exploration | 2026-05-04 | ShowModern data exploration tools often struggle to capture the subtleties of analytical intent, especially when users seek patterns that are difficult to specify using traditional query methods or natural language alone. We introduce a multimodal research probe for querying time-series and geospatial data that integrates free-form sketching, natural language, and visual annotations within a unified interaction space. Users articulate queries by sketching trends or spatial paths and augmenting them with annotations and analytical directives grounded in shared spatial and temporal context. The system employs a hybrid architecture combining geometric sketch matching and visual language models (VLMs) to support queries that interleave pattern matching and semantic constraints. Through a preliminary study with 20 participants, we observed recurring interaction patterns in which participants used spatial, temporal, and visual proximity to relate sketches, annotations, and language. Rather than treating these as isolated inputs, participants relied on their relative placement to disambiguate meaning. We analyze these behaviors as evidence for proximity semantics (PS), a form of deictic disambiguation in which meaning is shaped by the closeness of multimodal elements within a shared interaction space. We present PS as a conceptual lens grounded in observed user behavior, and discuss its implications for the design of future multimodal data exploration systems. | 5 pages, 5 figures |
| Simultaneous Inference for Nonlinear Time Series, a Sieve M-regression Approach | 2026-05-04 | ShowThis paper studies simultaneous inference of conditional distributions in nonlinear time series from a sieve M-regression perspective. Existing literature on sieve M-regression has primarily focused on pointwise asymptotics, leaving the development of uncertainty quantification over the entire predictor space unexplored. We address this gap by establishing a uniform Bahadur representation for the sieve M-estimator, accommodating dependent data and a growing number of sieve basis functions. A novel high-dimensional empirical process theory is developed for temporally dependent data, and a specifically designed M-decomposition method is utilized to control high-dimensional complexities. Building on this representation, we develop a convex Gaussian approximation to characterize the asymptotic behavior of the estimator and construct valid simultaneous confidence regions (SCRs). To facilitate practical implementation, we introduce a self-convolved bootstrap algorithm that accurately approximates the distribution of the maximal deviation. Our inferential framework is supported by rigorous error bounds and validated through numerical simulations and real data applications. | |
| RamanBench: A Large-Scale Benchmark for Machine Learning on Raman Spectroscopy | 2026-05-03 | ShowMachine Learning (ML) has transformed many scientific fields, yet key applications still lack standardized benchmarks. Raman spectroscopy, a widely used technique for non-invasive molecular analysis, is one such field where progress is limited by fragmented datasets, inconsistent evaluation, and models that fail to capture the structure of spectral data. We introduce RamanBench, the first large-scale, fully reproducible benchmark for ML on Raman spectroscopy, consisting of streamlined data access, evaluation protocols and code, as well as a live leaderboard. It unifies 74 datasets (including 16 first released with this benchmark) across four domains, comprising 325,668 spectra and spanning classification and regression tasks under diverse experimental conditions. We benchmark 28 models under a standardized protocol, including classical methods (e.g., PLS), Raman-specific (e.g., RamanNet), Tabular Foundation Model (TFM) (e.g., TabPFN), and time-series approaches (e.g., ROCKET). TFM consistently outperform domain-specific and gradient boosting baselines, while time-series models remain competitive. However, no method generalizes across datasets, revealing a fundamental gap. Therefore, we invite the community to contribute new approaches to our living benchmark, with the potential to accelerate advances in critical applications such as medical diagnostics, biological research, and materials science. | |
| FIRCE: A Framework for Intrusion Response and Conformal Evaluation | 2026-05-03 | ShowMachine learning-based intrusion detection systems deployed in real-world environments frequently suffer from model degradation due to concept drift, where changes in traffic patterns invalidate training assumptions. To address this, we present FIRCE, a Framework for Intrusion Response and Conformal Evaluation that augments supervised IDS classifiers with conformal evaluation-based uncertainty quantification and drift detection. FIRCE supports four conformal evaluation strategies: Inductive, Cross, Approximate Transductive, and our proposed Approximate Cross-Conformal Evaluator, which achieves robust performance with minimal calibration overhead. FIRCE also introduces an adaptive chunking mechanism that dynamically adjusts evaluation granularity in response to stream volatility, improving drift responsiveness while preserving computational efficiency. Using a custom IoT testbed of 10 commercial devices and time-series network captures under simulated attack and drift conditions, we demonstrate FIRCE's ability to detect distributional shifts and trigger model retraining. We additionally benchmark FIRCE on the CICIDS2018 and UNSW-NB15 datasets to validate its generalizability. Experimental results show that conformal evaluation-based drift detection, combined with adaptive chunking, enables an efficient and robust response to evolving threats. | 9 pag...9 pages, 4 algorithms, 3 tables, 1 figure, accepted by IEEE SmartNets26 |
| ASTER: Latent Pseudo-Anomaly Generation for Unsupervised Time-Series Anomaly Detection | 2026-05-03 | ShowTime-series anomaly detection (TSAD) is critical in domains such as industrial monitoring, healthcare, and cybersecurity, but it remains challenging due to rare and heterogeneous anomalies and the scarcity of labelled data. This scarcity makes unsupervised approaches predominant, yet existing methods often rely on reconstruction or forecasting, which struggle with complex data, or on embedding-based approaches that require domain-specific anomaly synthesis and fixed distance metrics. We propose ASTER, a framework that generates pseudo-anomalies directly in the latent space, avoiding handcrafted anomaly injections and the need for domain expertise. A latent-space decoder produces tailored pseudo-anomalies to train a Transformer-based anomaly classifier, while a pre-trained LLM enriches the temporal and contextual representations of this space. Experiments on three benchmark datasets show that ASTER achieves state-of-the-art performance and sets a new standard for LLM-based TSAD. | |
| Applying the Polynomial Maximization Method to Estimate ARIMA Models with Asymmetric Non-Gaussian Innovations | 2026-05-03 | ShowClassical estimators for ARIMA parameters (MLE, CSS, OLS) assume Gaussian innovations, an assumption frequently violated in financial and economic data exhibiting asymmetric distributions with heavy tails. We develop and validate the second-order polynomial maximization method (PMM2) for estimating ARIMA$(p,d,q)$ models with non-Gaussian innovations. PMM2 is a semiparametric technique that exploits higher-order moments and cumulants without requiring full distributional specification. Monte Carlo experiments (128,000 simulations) across sample sizes $N \in {100, 200, 500, 1000}$ and four innovation distributions demonstrate that PMM2 substantially outperforms classical methods for asymmetric innovations. For ARIMA(1,1,0) with $N=500$, relative efficiency reaches 1.58--1.90 for Gamma, lognormal, and $χ^2(3)$ innovations (37--47% variance reduction). Under Gaussian innovations PMM2 matches OLS efficiency, avoiding the precision loss typical of robust estimators. The method delivers major gains for moderate asymmetry ($ | γ_3 |
| Persistent Homology of Time Series through Complex Networks | 2026-05-02 | ShowWe present a unified pipeline for univariate time series classification via complex networks and persistent homology. A time series is mapped to a graph through one of five constructions across three families (visibility (natural and horizontal visibility graphs), transition, and proximity) and the graph is converted to a dissimilarity matrix from which a Vietoris-Rips filtration yields persistence diagrams. These diagrams are vectorized into fixed-length features through persistence landscapes and topological summary statistics. By standardizing the downstream processing, differences in classification performance are attributable to the network construction and distance metric alone. Experiments on twelve UCR benchmarks show that (i) no single construction dominates: the optimal graph type depends on the signal's discriminative structure; (ii) the graph distance metric is a first-order design choice, with diffusion distance uniformly outperforming shortest-path alternatives; and (iii) persistence-based features degrade gracefully under noise, consistent with the classical stability theorem of persistent homology. | |
| Why Model Selection Fails in Time Series Forecasting: An Empirical Study of Instability Across Data Regimes | 2026-05-02 | ShowTime series forecasting models often exhibit inconsistent performance across datasets with varying statistical and structural properties. Despite the wide range of available forecasting techniques, it remains unclear whether model selection can be reliably guided by simple data characteristics. This paper investigates why rule-based model selection fails in time series forecasting by analyzing the relationship between data-regime descriptors and model performance. A descriptor-based framework is introduced to characterize time series using measurable properties, including trend strength, seasonality, noise level, and temporal dependence. Based on these descriptors, a rule-based selection mechanism is formulated to map data regimes to candidate forecasting models. The approach is evaluated on multiple real-world datasets across different domains and forecasting horizons. The results show that rule-based model selection achieves low accuracy, with correct model identification occurring in only a small fraction of cases. Significant discrepancies are observed between recommended and empirically optimal models, particularly in noisy and mixed regimes. Further analysis reveals that model performance is highly sensitive to both dataset characteristics and forecasting horizon, resulting in substantial ranking instability across scenarios. These findings explain why simple heuristic rules fail to generalize and demonstrate that forecasting performance cannot be reliably predicted using static, descriptor-based approaches. This study provides empirical evidence that model selection in time series forecasting is inherently context-dependent and highlights the need for more adaptive, data-driven strategies. | |
| Ellipsoidal Time Series Forecasting | 2026-05-02 | ShowWe argue that long-term forecasting requires learning local Jacobians with explicit spectral structure, going beyond simple conditional mean matching. Our method, Fern, invokes Brenier's theorem to directly parameterize the Jacobian as a symmetric positive semi-definite (SPD) factorization, treating forecasting as the optimal transport of probability mass from a fixed Gaussian source to data-dependent ellipsoids. This formulation reduces the computational cost of eigendecomposition from cubic to linear time while providing interpretable, geometry-aware projections. To rigorously evaluate robustness, we introduce a synthetic benchmark with controlled non-stationary shocks alongside new metrics like Effective Prediction Time (EPT). Fern demonstrates exceptional stability, outperforming baselines like DLinear and Koopa by over two orders of magnitude (up to 790x) on nonstationary settings where standard benchmarks fail to expose model brittleness. | Accep...Accepted by ICML 2026. Public code at https://anonymous.4open.science/r/ FernPaper-58B4 |
| TimeTok: Granularity-Controllable Time-Series Generation via Hierarchical Tokenization | 2026-05-02 | ShowTime-series generative models often lack control over temporal granularity, forcing users to accept whatever granularity the model produces. To enable truly user-driven generation, we introduce TimeTok, a unified framework for Granularity-Controllable Time-Series Generation (GC-TSG), which generates time series at any target granularity from any coarser input (e.g., rough sketches) or from scratch. At the core of TimeTok is a hierarchical tokenization strategy that maps time series into an ordered sequence of tokens, from coarse to fine temporal granularity. Our autoregressive generation process operates across these granularity levels, producing token blocks that are decoded back into continuous time series. This design naturally enables GC-TSG - including standard generation - within a single framework, where controlling the number of token blocks provides explicit control over output detail. Experiments show that TimeTok excels at GC-TSG tasks while achieving state-of-the-art performance in standard generation. Furthermore, we showcase TimeTok's potential as a foundational tokenizer by training on multiple datasets with heterogeneous temporal granularities, verifying strong transferability that consistently outperforms models trained on individual datasets. To our knowledge, this is the first unified framework that covers the full generative spectrum for time series, offering a valuable foundation for models that benefit from diverse temporal granularities. | |
| Toward a foundational thermal model for residential buildings | 2026-05-02 | ShowThe building energy community lacks a foundational thermal model, i.e., a single pretrained model capable of generalizing across diverse buildings, climates, and control strategies without building-specific calibration. Achieving this vision requires architectural principles that capture universal thermal dynamics rather than memorizing building-specific patterns. We take a step toward this goal by presenting a physics-informed transformer architecture that embeds domain knowledge, e.g., derivative enrichment and Euler-based numerical integration, into a decoder-only framework. We incorporate static building features extracted from simulation models and employ Rotary Position Embedding attention to capture temporal dependencies. Evaluated on the CityLearn dataset spanning 247 residential buildings across three climate zones, our model achieves one-step prediction accuracy (RMSE of 0.30°C in Texas, 0.29°C in Vermont) while outperforming both traditional baselines and fine-tuned Time-Series Foundation Models. We also demonstrate zero-shot transferability: models trained on as few as two buildings generalize to unseen buildings and climate zones without fine-tuning. Despite the limitation of simulated residential buildings, our results establish physics-informed architectural principles as a promising foundation for universal building thermal models. | |
| Time-series forecasting through the lens of dynamics | 2026-05-02 | ShowWhile deep learning is facing an homogenization across modalities led by Transformers, they are still challenged by shallow linear models in the time-series forecasting task. Our hypothesis is that models should learn a direct link from past to future data points, which we identify as a learning dynamics capability. We develop an original $\texttt{PRO-DYN}$ nomenclature to analyze existing models through the lens of dynamics. Two observations thus emerge: $\textbf{1.}$ under-performing architectures learn dynamics at most partially, $\textbf{2.}$ the location of the dynamics block at the model end is of prime importance. Our systemic and empirical studies both confirm our observations on a set of performance-varying models with diverse backbones. We propose a simple plug-and-play methodology guiding model designs and improvements. | Accep...Accepted at ICML 2026 |
| Survival Analysis with Machine Learning for Predicting Li-ion Battery Remaining Useful Life | 2026-05-02 | ShowBattery degradation significantly impacts the reliability and efficiency of energy storage systems, particularly in electric vehicles and industrial applications. Predicting the remaining useful life (RUL) of lithium-ion batteries is crucial for optimizing maintenance schedules, reducing costs, and improving safety. Traditional RUL prediction methods often struggle with nonlinear degradation patterns and uncertainty quantification. To address these challenges, we propose a hybrid survival analysis framework integrating survival data reconstruction, survival model learning, and survival probability estimation. Our approach transforms battery voltage time series into time-to-failure data using path signatures. The multiple Cox-based survival models and machine-learning-based methods, such as DeepHit and MTLR, are learned to predict battery failure-free probabilities over time. Experiments conducted on the Toyota battery and NASA battery datasets demonstrate the effectiveness of our approach, achieving high time-dependent AUC and concordance index (C-Index) while maintaining a low integrated Brier score. The data and source codes are available to the public at https://github.com/okic-ai/rul | |
| Factor State Space Modelling of the Ornstein-Uhlenbeck Process with Measurement Error and its Application | 2026-05-02 | ShowStandard Ornstein-Uhlenbeck (OU) models often yield biased parameter estimates when measurement error is ignored. While the Ornstein-Uhlenbeck State Space Model (OUSSM) addresses this in univariate settings, multidimensional extensions remain limited. This paper introduces the factor OUSSM to model multi-dimensional, mean-reverting systems with observational noise. We resolve critical identifiability challenges in parameter estimation by establishing necessary constraints and validating the method through extensive simulations. We demonstrate the model's versatility by analyzing human gut microbiome dynamics and North Atlantic Sea Surface Temperature (SST) data. The results reveal distinct latent temporal structures in both biological and environmental systems, establishing the factor OUSSM as a robust framework for multivariate time series analysis. | |
| CombinationTS: A Modular Framework for Understanding Time-Series Forecasting Models | 2026-05-02 | ShowRecent progress in time-series forecasting has led to rapidly increasing architectural complexity, yet many reported State-of-the-Art gains are statistically fragile or misattributed. We argue that progress requires a shift from model selection to modular attribution, identifying which components truly drive performance. We propose CombinationTS, a self-contained probabilistic evaluation framework that decomposes forecasting models into orthogonal modules--Input Transformation, Embedding, Encoder, Decoder, and Output Transformation--and evaluates them under a shared evaluation condition space. By quantifying each component via marginalized performance ($μ$) and stability ($σ$), CombinationTS enables robust attribution beyond fragile point estimates. Through large-scale paired evaluation, we uncover the Identity Paradox: once the data view (Embedding) is well-designed, a parameter-free Identity Encoder often matches or outperforms complex backbones. We further show that explicit structural priors introduced via Input Transformations yield a more favorable performance-stability trade-off than increasing Encoder complexity, establishing a principled baseline for architectural necessity. | Accep...Accepted by ICML 2026 main track. Code available at https://github.com/BenchCouncil/CombinationTS |
| Capture Timing-Attention of Events in Clinical Time Series | 2026-05-01 | ShowAutomatically discovering personalized sequential events from large-scale time-series data is crucial for enabling precision medicine in clinical research, yet it remains a formidable challenge even for contemporary AI models. For example, while transformers capture rich associations, they are mostly agnostic to event timing and ordering, thereby bypassing potential causal reasoning. Intuitively, we need a method capable of evaluating the "degree of alignment" among patient-specific trajectories and identifying their shared patterns, i.e., the significant events in a consistent sequence. This necessitates treating timing as a true \emph{computable} dimension, allowing models to assign relative timestamps'' to candidate events beyond their observed physical times. In this work, we introduce LITT (Individual-Level Time Transformation), a novel architecture that enables temporary alignment of sequential events on a virtual relative timeline'', thereby enabling \emph{event-timing-focused attention} and personalized interpretations of clinical trajectories. Its interpretability and effectiveness are validated on real-world longitudinal EHR data from 3,276 breast cancer patients to predict the onset timing of cardiotoxicity-induced heart disease. Furthermore, LITT outperforms both the benchmark and state-of-the-art survival analysis methods on public datasets, positioning it as a significant step forward for precision medicine in clinical AI. |
8 pages of body text |
| MambaSL: Exploring Single-Layer Mamba for Time Series Classification | 2026-05-01 | ShowDespite recent advances in state space models (SSMs) such as Mamba across various sequence domains, research on their standalone capacity for time series classification (TSC) has remained limited. We propose MambaSL, a framework that minimally redesigns the selective SSM and projection layers of a single-layer Mamba, guided by four TSC-specific hypotheses. To address benchmarking limitations -- restricted configurations, partial University of East Anglia (UEA) dataset coverage, and insufficiently reproducible setups -- we re-evaluate 20 strong baselines across all 30 UEA datasets under a unified protocol. As a result, MambaSL achieves state-of-the-art performance with statistically significant average improvements, while ensuring reproducibility via public checkpoints for all evaluated models. Together with visualizations, these results demonstrate the potential of Mamba-based architectures as a TSC backbone. | accep...accepted at ICLR 2026 |
| Pi-Change: A Prior-Informed Multiple Change Point Detection Algorithm | 2026-05-01 | ShowStatistical change point (CP) detection methods typically rely on likelihood-based inference and ignore contextual information about plausible CP locations beyond the observed sequence. Although informative priors provide a natural way to incorporate such information, general and computationally efficient methods for doing so are lacking, especially for multiple CP detection. To address this gap, we propose a prior-informed CP detection algorithm (Pi-Change) that incorporates prior information on CP locations through a time-varying penalty term. We prove that the proposed penalty can be embedded in the Pruned Exact Linear Time framework while preserving the dynamic programming recursion and pruning rule required for efficient multiple CP detection. Across simulation studies and three time-series applications, Pi-Change discourages spurious CPs unsupported by prior information, remains robust to prior misspecification, and improves detection accuracy. More broadly, Pi-Change extends multiple CP detection beyond purely data-driven fitting by incorporating partial prior knowledge in a computationally efficient and interpretable way. It is particularly useful when CPs arise from heterogeneous mechanisms or are associated with known external events, helping quantify the delay between an event and the resulting structural change. | |
| From Prediction to Practice: A Task-Aware Evaluation Framework for Blood Glucose Forecasting | 2026-05-01 | ShowClinical time-series forecasting is increasingly studied for decision support, yet standard aggregate metrics can obscure whether a model is actually useful for the task it is meant to serve. In safety-critical settings, low average error can coexist with dangerous failures in exactly the high-risk regimes that matter most. We present a task-aware evaluation framework for blood glucose forecasting built around two downstream uses: hypoglycemia early warning and insulin dosing decision support. For early warning, we evaluate on real data from three clinical cohorts using event-level recall and false alarms per patient-day, metrics that reflect operational alarm burden rather than aggregate accuracy. We show that models appearing acceptable overall, with recall above 0.9 on the full test set, can fail badly in the post-bolus slice, where insulin-on-board is elevated and missed warnings carry the greatest clinical consequences. Standard forecasting evaluation, however, does not test whether a model can reason about the effects of actions, a requirement for supporting insulin dosing decisions. We therefore add a second, interventional arm using the FDA-accepted UVA/Padova simulator, where we evaluate whether forecasters can predict glucose responses to altered insulin plans in paired factual/counterfactual scenarios. We show that models that look strong on real-data forecasting often fail to predict the direction, magnitude, or ranking of intervention effects, and choose poor insulin doses when evaluated under a clinically motivated cost. Taken together, the two arms reveal a consistent gap between forecasting accuracy and task-relevant usefulness. We release the benchmark, the standardized preprocessing pipeline for public cohorts, and the simulator-based interventional dataset as a reproducible toolkit. | |
| GCGNet: Graph-Consistent Generative Network for Time Series Forecasting with Exogenous Variables | 2026-05-01 | ShowExogenous variables offer valuable supplementary information for predicting future endogenous variables. Forecasting with exogenous variables needs to consider both past-to-future dependencies (i.e., temporal correlations) and the influence of exogenous variables on endogenous variables (i.e., channel correlations). This is pivotal when future exogenous variables are available, because they may directly affect the future endogenous variables. Many methods have been proposed for time series forecasting with exogenous variables, focusing on modeling temporal and channel correlations. However, most of them use a two-step strategy, modeling temporal and channel correlations separately, which limits their ability to capture joint correlations across time and channels. Furthermore, in real-world scenarios, time series are frequently affected by various forms of noises, underscoring the critical importance of robustness in such correlations modeling. To address these limitations, we propose GCGNet, a Graph-Consistent Generative Network for time series forecasting with exogenous variables. Specifically, GCGNet first employs a Variational Generator to produce coarse predictions. A Graph Structure Aligner then further guides it by evaluating the consistency between the generated and true correlations, where the correlations are represented as graphs, and are robust to noises. Finally, a Graph Refiner is proposed to refine the predictions to prevent degeneration and improve accuracy. Extensive experiments on 12 real-world datasets demonstrate that GCGNet outperforms state-of-the-art baselines. | |
| Scalable Context-Aware Graph Attention for Unsupervised Anomaly Detection in Large-Scale Mobile Networks | 2026-05-01 | ShowMobile network operators must monitor thousands of heterogeneous network elements across the radio access network and the packet core, each exposing high-dimensional KPI time series. The scale and cost of incident labelling make supervised approaches impractical, motivating unsupervised anomaly detection robust to context shifts and nonstationarity. We propose \textbf{C-MTAD-GAT} (\emph{Context-aware Multivariate Time-series Anomaly Detection with Graph Attention}), an anomaly detection framework designed to operate as a single shared model across large populations of network elements. The model combines temporal and feature-wise graph attention with lightweight static and dynamic context conditioning and a dual-head decoder for reconstruction and multi-step forecasting. It produces per-element, per-feature anomaly scores, converted to alerts via fully unsupervised thresholds calibrated from validation residuals. On the TELCO dataset released with DC-VAE \cite{garcia2023onemodel}, C-MTAD-GAT improves event-level affiliation and pointwise F1 while generating fewer alarms than prior graph-attention and VAE-based baselines. We then apply the same system to nation-scale radio access and evolved packet core control-plane counter data from a mobile network operator, where it is deployed. Operator feedback indicates the alerts are actionable and support daily monitoring, showing scalability across domains without relying on labelled incidents. | This ...This work has been submitted to the IEEE for possible publication |
| PAMod: Modeling Cyclical Shifts via Phase-Amplitude Modulation for Non-stationary Time Series Forecasting | 2026-05-01 | ShowReal-world time series forecasting faces the fundamental challenge of non-stationary statistical properties, including shifts in mean and variance over time. While reversible instance normalization (RevIN) has shown promise by stationarizing inputs and denormalizing outputs, it relies on the strong assumption that historical and future distributions remain identical. We observe that in many practical applications, distribution shifts follow cyclical patterns that correlate with periodic positions (e.g., seasonal and holiday volatility). To this end, we propose PAMod, a lightweight yet powerful framework that models cyclical distribution shifts via Phase-Amplitude Modulation in the normalized feature space. PAMod learns periodic embeddings to modulate representations: phase modulation captures mean shifts, while amplitude modulation adapts to variance changes. Crucially, we prove mathematically that modulating in normalized space is equivalent to applying dynamic denormalization, offering an elegant unification of distribution adaptation and representation learning. Extensive experiments on twelve real-world benchmarks demonstrate that PAMod achieves state-of-the-art performance with fewer computational resources. Furthermore, our modulation mechanism, as a novel plug-and-play technique, can improve existing time-series forecasting methods with simple integration. | |
| PAMNet: Cycle-aware Phase-Amplitude Modulation Network for Multivariate Time Series Forecasting | 2026-05-01 | ShowReliable periodic patterns serve as a fundamental basis for accurate multivariate time series forecasting. However, existing methods either implicitly extract periodicity through complex model architectures (e.g., Transformers) with high computational overhead or overlook the intrinsic phase-amplitude coupling when modeling periodic components explicitly. To address these issues, we propose a novel Cycle-aware Phase-Amplitude Modulation Network (PAMNet) that explicitly decomposes periodic patterns into complementary phase and amplitude components. The core innovation lies in its dual-branch modulator, featuring dedicated learnable embeddings for phase positioning and amplitude modulation. The phase branch employs cyclical embeddings to capture phase-dependent mean shifts, while the amplitude branch models intensity variations to adapt to changes in variance. A lightweight modulator with element-wise fusion efficiently combines these components, enabling explicit modeling of their interactions without complex attention mechanisms. Extensive experiments on twelve real-world datasets demonstrate that our method achieves state-of-the-art performance through its novel phase-amplitude decoupling mechanism, offering a new perspective for cyclical modeling in time series forecasting. | |
| Characterizing control between interacting subsystems with deep Jacobian estimation | 2026-05-01 | ShowBiological function arises through the dynamical interactions of multiple subsystems, including those between brain areas, within gene regulatory networks, and more. A common approach to understanding these systems is to model the dynamics of each subsystem and characterize communication between them. An alternative approach is through the lens of control theory: how the subsystems control one another. This approach involves inferring the directionality, strength, and contextual modulation of control between subsystems. However, methods for understanding subsystem control are typically linear and cannot adequately describe the rich contextual effects enabled by nonlinear complex systems. To bridge this gap, we devise a data-driven nonlinear control-theoretic framework to characterize subsystem interactions via the Jacobian of the dynamics. We address the challenge of learning Jacobians from time-series data by proposing the JacobianODE, a deep learning method that leverages properties of the Jacobian to directly estimate it for arbitrary dynamical systems from data alone. We show that JacobianODEs outperform existing Jacobian estimation methods on challenging systems, including high-dimensional chaos. Applying our approach to a multi-area recurrent neural network (RNN) trained on a working memory selection task, we show that the "sensory" area gains greater control over the "cognitive" area over learning. Furthermore, we leverage the JacobianODE to directly control the trained RNN, enabling precise manipulation of its behavior. Our work lays the foundation for a theoretically grounded and data-driven understanding of interactions among biological subsystems. | 10 pages, 6 figures |
| GaMMA: Towards Joint Global-Temporal Music Understanding in Large Multimodal Models | 2026-05-01 | ShowIn this paper, we propose GaMMA, a state-of-the-art (SoTA) large multimodal model (LMM) designed to achieve comprehensive musical content understanding. GaMMA inherits the streamlined encoder-decoder design of LLaVA, enabling effective cross-modal learning between music and language. By incorporating audio encoders in a mixture-of-experts manner, GaMMA effectively unifies both time-series and non-time-series music understanding tasks within one set of parameters. Our approach combines carefully curated datasets at scale with a progressive training pipeline, effectively pushing the boundaries of music understanding via pretraining, supervised fine-tuning (SFT), and reinforcement learning (RL). To comprehensively assess both temporal and non-temporal capability of music LMMs, we introduce MusicBench, the largest music-oriented benchmark, comprising 3,739 human-curated multiple-choice questions covering diverse aspects of musical understanding. Extensive experiments demonstrate that GaMMA establishes new SoTA in the music domain, achieving 79.1% accuracy on MuchoMusic, 79.3% on MusicBench-Temporal, and 81.3% on MusicBench-Global, consistently outperforming previous methods. | |
| Time-series Meets Complex Motion Modeling: Robust and Computational-effective Motion Predictor for Multi-object Tracking | 2026-05-01 | ShowMulti-object tracking (MOT) is critical in numerous real-world applications, including surveillance, autonomous driving, and robotics. Accurately predicting object motion is fundamental to MOT, but current methods struggle with the complexities of real-world, non-linear motion (e.g., sudden stops, sharp turns). While recent research has gravitated towards increasingly complex and computationally expensive generative models to tackle this problem, their practical utility is often constrained. This paper challenges that paradigm, arguing that such complexity is not only unnecessary but can be outperformed by a more efficient, purpose-built approach. We introduce the Temporal Convolutional Motion Predictor (TCMP), a novel framework for MOT that leverages a modified Temporal Convolutional Network (TCN) featuring dilated convolutions and a regression head. This design allows for effective motion prediction across arbitrary temporal context lengths. Experimental results demonstrate that our approach achieves state-of-the-art performance, specifically improves upon the previous best method in several key metrics: HOTA (a measure of overall tracking accuracy) increases from 62.3% to 63.4%, IDF1 (a measure of identity preservation) rises from 63.0% to 65.0%, and AssA (a measure of association accuracy) improves from 47.2% to 49.1%. Significantly, TCMP achieves this performance while being highly efficient; it has only 0.014 times the parameters and requires only 0.05 times the computational cost (FLOPs) compared to the SOTA method. while is only 0.014 times the size (in terms of parameters) and requires only 0.05 times the computational cost (in terms of FLOPs). These findings highlight the robustness of our method to advance MOT systems by ensuring adaptability, accuracy, and efficiency in complex tracking environments. | |
| CGM-JEPA: Learning Consistent Continuous Glucose Monitor Representations via Predictive Self-Supervised Pretraining | 2026-05-01 | ShowContinuous Glucose Monitoring (CGM) can detect early metabolic subphenotypes (insulin resistance, IR; $β$-cell dysfunction), but population-scale deployment faces two coupled problems. First, the same physiological state appears through multiple views (CGM time series, venous OGTT, Glucodensity summaries), so single-view representations fail to transfer when deployment shifts the modality or setting. Second, baselines perform inconsistently across these shifts. Both problems point to one remedy: representations that abstract away from any single view to capture higher-level temporal and distributional structure. We propose CGM-JEPA, a self-supervised pretraining framework which predicts masked latent representations rather than raw values, yielding abstraction that transfers across modalities. X-CGM-JEPA adds a masked Glucodensity cross-view objective for complementary distributional information. We pretrain on $\sim$389k unlabeled CGM readings from 228 subjects and evaluate on two clinical cohorts ($N=27$ and $N=17$ public-release subsets) across three regimes (cohort generalization, venous-to-CGM transfer, home CGM) under 20-iteration $\times$ 2-fold cross-validation. X-CGM-JEPA ranks first or second on AUROC for both endpoints across all three regimes while no baseline does, exceeding the strongest baseline by up to $+6.5$ pp in cohort generalization and $+3.6$ pp in venous-to-CGM transfer (paired Wilcoxon, $p<0.001$). Under modality shift, it matches mean AUROC while redistributing toward weaker subgroups (ethnicity AUROC gap shrinks 25-54%); on sparse in-domain venous data, the distributional view lifts label-aware clustering (ARI $+39%$, NMI $+40%$). Code and weights: https://github.com/cruiseresearchgroup/CGM-JEPA | |
| Efficient Spatio-Temporal Vegetation Pixel Classification with Vision Transformers | 2026-04-30 | ShowPlant phenology-the study of recurrent life cycle events-is essential for understanding ecosystem dynamics and their responses to climate change impacts. While Unmanned Aerial Vehicles (UAVs) and near-surface cameras enable high-resolution monitoring, identifying plant species across time remains computationally challenging. State-of-the-art approaches, specifically Multi-Temporal Convolutional Networks (CNNs), rely on rigid multi-branch architectures that scale poorly with longer time series and require large spatial context windows. In this paper, we present an extensive study on optimizing Vision Transformers (ViTs) for efficient spatio-temporal vegetation pixel classification. We conducted a comprehensive ablation study analyzing seven key design dimensions, including: (i) data normalization; (ii) spectral arrangement; (iii) boundary handling; (iv) spatial context window shape and size; (v) tokenization strategies; (vi) positional encoding; and (vii) feature aggregation strategies. Our method was evaluated on two datasets from the Brazilian Cerrado biome, Serra do Cipó (aerial imagery) and Itirapina (near-surface imagery). Experimental results demonstrate that our ViT approach offers a substantial improvement in computational efficiency while maintaining competitive classification performance. Notably, our ViT reduces Floating Point Operations (FLOPs) by an order of magnitude and maintains constant parameter complexity regardless of the time series length, whereas the CNN baseline scales linearly. Our findings confirm that ViTs are a robust, scalable solution for resource-constrained phenological monitoring systems. | |
| Training-Free Time Series Classification via In-Context Reasoning with LLM Agents | 2026-04-30 | ShowTime series classification (TSC) spans diverse application scenarios, yet labeled data are often scarce, making task-specific training costly and inflexible. Recent reasoning-oriented large language models (LLMs) show promise in understanding temporal patterns, but purely zero-shot usage remains suboptimal. We propose FETA, a multi-agent framework for training-free TSC via exemplar-based in-context reasoning. FETA decomposes a multivariate series into channel-wise subproblems, retrieves a few structurally similar labeled examples for each channel, and leverages a reasoning LLM to compare the query against these exemplars, producing channel-level labels with self-assessed confidences; a confidence-weighted aggregator then fuses all channel decisions. This design eliminates the need for pretraining or fine-tuning, improves efficiency by pruning irrelevant channels and controlling input length, and enhances interpretability through exemplar grounding and confidence estimation. On nine challenging UEA datasets, FETA achieves strong accuracy under a fully training-free setting, surpassing multiple trained baselines. These results demonstrate that a multi-agent in-context reasoning framework can transform LLMs into competitive, plug-and-play TSC solvers without any parameter training. The code is available at https://github.com/SongyuanSui/FETATSC. | 8 pag...8 pages main content, 12 pages total including appendix, 1 figure |
| Selfie-Capture Dynamics as an Auxiliary Signal Against Deepfakes and Injection Attacks for Mobile Identity Verification | 2026-04-30 | ShowMobile remote identity verification (RIdV) systems are exposed to attacks that manipulate or replace the facial video stream, including presentation attacks, real-time deepfakes, and video injection. Recent European requirements, including ETSI TS 119 461 and CEN/TS 18099, motivate complementary evidence channels beyond camera-based presentation-attack detection. This paper investigates whether passive motion traces recorded during selfie capture provide auxiliary evidence for spoof screening and user verification. We introduce CanSelfie, a dataset of 375 bona fide multi-sensor sequences collected at 50,Hz from 30 participants using a commercial mobile RIdV application, together with stationary, handheld, and temporally shifted attack-proxy scenarios. We benchmark 7 multivariate time-series classifiers and 8 whole-series anomaly detectors across sensor configurations and temporal windows. For spoof screening, accelerometer-only ROCKAD obtains 0.00% false rejection rate (FRR) and 43.8% false acceptance rate (FAR), while QUANT+3-NN obtains the lowest overall FAR of 32.0% at 2.37% FRR; both reject all stationary attack proxies. For same-device and same-session user verification, WEASEL+MUSE reaches 1.07% equal error rate (EER) using 9 sensor channels. The analysis shows that raw accelerometer data, preserving gravity and orientation cues, is the most informative modality, and that closed-set classification accuracy alone does not imply good verification performance because threshold calibration depends on score distributions. The findings suggest that short selfie-capture motion traces contain measurable spoof-related and identity-related information, supporting their use as a low-friction auxiliary signal while also identifying the need for cross-device, cross-session, and real injection-attack evaluation. | 12 pa...12 pages, 5 figures, 8 tables, 51 references, conference |
| PhaseNet++: Phase-Aware Frequency-Domain Anomaly Detection for Industrial Control Systems via Phase Coherence Graphs | 2026-04-30 | ShowMultivariate time series anomaly detection in ICS has attracted growing attention due to the increasing threat of cyber-physical attacks on critical infrastructure. State-of-the-art methods model inter-sensor relationships from raw time-domain amplitude values, using graph neural networks, Transformers. However, these methods discard the phase spectrum produced by time frequency transformations, We argue that phase information constitutes a complementary and previously overlooked detection modality for ICS anomaly detection. We present PhaseNet++, a frequency-domain autoencoder that operates on the Short-Time Fourier Transform (STFT) of sliding sensor windows, retaining both magnitude and phase spectra. A Phase Coherence Index (PCI), inspired by the Phase Locking Value from neuroscience, summarizes pairwise phase consistency across frequency bins into a continuous adjacency matrix. This matrix guides a graph attention network that propagates information preferentially among phase-synchronized sensors. A sensor-token Transformer encoder captures system-wide structure, and a dual-head decoder reconstructs magnitude and phase jointly via circular and coherence-aware objectives. Evaluated on the Secure Water Treatment (SWaT) benchmark, PhaseNet++ achieves an F1-score of 90.98%, ROC-AUC of 95.66%, and average precision of 91.51%. Ablation studies show that the phase-aware front-end and PCI graph module together add only 264,816 parameters, demonstrating that the phase inductive bias is lightweight. While the absolute F1-score is second best than that of all recent raw-value methods evaluated under different protocols, we position this work as the first systematic study of phase-domain anomaly detection for ICS. | 9 pages, 1 figure |
| SHIFT: Robust Double Machine Learning for Average Dose-Response Functions under Heavy-Tailed Contamination | 2026-04-30 | ShowDouble-machine-learning pipelines for the Average Dose-Response Function rely on kernel-weighted local-linear smoothers, which inherit unbounded functional influence: a single outlier within a kernel window biases the curve across the entire window. We introduce SHIFT (Self-calibrated Heavy-tail Inlier-Fit with Tempering), a robust DML estimator combining cross-fit nuisance orthogonalization with a kernel-local Welsch-loss second stage optimized by Graduated Non-Convexity, and -- the principal design choice -- a defensive OLS refit whose inlier cutoff is scaled by post-GNC residual MAD rather than the raw-outcome MAD. On a localized-contamination stress test at $p=0.25$ this design choice drops level-RMSE from 1.03 to 0.33 while leaving clean and uniformly-contaminated runs unchanged. Across 1,400 main-sweep fits, SHIFT has competitive worst-case shape recovery (RMSE $0.325$ at $p=0.25$, second to Huber-DML's $0.276$); among the three methods with worst-case RMSE below $0.35$, only SHIFT emits a non-uniform per-sample weight vector, recovering the ground-truth outlier mask at mean $F_1 \approx 0.96$ (range $0.945$--$0.968$) on Gaussian-jump DGPs. We pair the estimator with a six-technique Extreme Value Theory diagnostic suite (Hill, GPD-MLE/PWM, GEV, Mean Excess, parameter stability, causal tail coefficient) that lets a practitioner distinguish Frechet from Weibull regimes and choose between SHIFT and L1 alternatives on empirical grounds. Extensions to binary-treatment CATE (Huber pseudo-outcome X-Learner) and time-series ADRF (block-CV + rolling MAD) are included. A counter-intuitive ablation: linear nuisance models (Ridge, Lasso) outperform gradient-boosted nuisances for robust DML under uniform contamination, inverting the usual more-flexible-is-better heuristic. | 77 pa...77 pages, 43 figures, 35 tables. Code and raw CSVs: https://github.com/EichiUehara/ADRF-Robust-DML |
| CaTS-Bench: Can Language Models Describe Time Series? | 2026-04-30 | ShowTime series captioning, the task of describing time series in natural language, requires numeric and temporal reasoning, trend interpretation, and contextual understanding. Existing benchmarks, however, often rely on fully synthetic or generic captions, and typically neglect metadata and visual representations. We introduce CaTS-Bench, a comprehensive benchmark for Context-aware Time Series reasoning across 11 diverse domains, centered on a gold-standard evaluation set of 1746 human-rewritten captions that measure how effectively models translate numeric trends into immediately interpretable narratives. To address the scarcity of human-annotated data, we also propose a scalable pipeline for generating high-fidelity synthetic captions, the quality of which we validate. We evaluate leading Vision-Language Models on our benchmark, revealing that even proprietary models struggle to capture numeric nuances in temporal descriptions, while finetuning open-source models on synthetic data yields substantial performance gains. Finally, we release a diagnostic suite of 910 multiple-choice questions and use tailored numeric metrics to gauge time-series-specific reasoning capabilities, establishing CaTS-Bench as a reliable foundation for grounded, multimodal text generation in numeric domains. | 9 pag...9 pages, 6 figures, 4 tables in the main paper. Many more in the appendix |
| ARFBench: Benchmarking Time Series Question Answering Ability for Software Incident Response | 2026-04-30 | ShowTime series question-answering (TSQA), in which we ask natural language questions to infer and reason about properties of time series, is a promising yet underexplored capability of foundation models. In this work, we present ARFBench, a TSQA benchmark that evaluates the understanding of multimodal foundation models (FMs) on time series anomalies prevalent in software incident data. ARFBench consists of 750 questions across 142 time series and 5.38M data points from 63 production incidents sourced exclusively from internal telemetry at Datadog. We evaluate leading proprietary and open-source LLMs, VLMs, and time series FMs and observe that frontier VLMs perform markedly better than existing baselines; the leading model (GPT-5) achieves a 62.7% accuracy and 51.9% F1. We next demonstrate the promise of specialized multimodal approaches. We develop a novel TSFM + VLM hybrid prototype which we post-train on a small set of synthetic and real data that yields comparable overall F1 and accuracy with frontier models. Lastly, we find models and human domain experts exhibit complementary strengths. We define a model-expert oracle, a best-of-2 oracle selector over model and expert answers, yielding 82.8% F1 and 87.2% accuracy and establishing a new superhuman frontier for future TSQA models. The benchmark is available at https://huggingface.co/datasets/Datadog/ARFBench. | Updat...Updated author affiliation |
| Learning Fingerprints for Medical Time Series with Redundancy-Constrained Information Maximization | 2026-04-30 | ShowLearning meaningful representations from medical time series (MedTS) such as ECG or EEG signals is a critical challenge. These signals are often high-dimensional, variable-length and rife with noise. Existing self-supervised approaches, such as Masked Autoencoders (MAEs) are highly effective for pre-training general-purpose encoders. However, they do not explicitly learn compact and semantically interpretable latent representations, typically relying on heuristic aggregation strategies such as global average pooling or a designated [CLS] token. We propose a novel framework that compresses a variable-length MedTS into a fixed-size set of $k$ latent Fingerprint Tokens. Our architecture employs a cross-attention bottleneck to generate these tokens and is trained with a dual-objective function. The first objective is a reconstruction loss, which ensures the tokens are \textit{sufficient statistics} for the original data. The second, a diversity penalty based on the Total Coding Rate (TCR), explicitly minimizes the redundancy between tokens, encouraging them to become statistically \textit{disentangled} representations. We present the theoretical justification for our method, framing it as a novel \textbf{Disentangled Rate-Distortion} problem. This approach produces a low-dimensional, interpretable, and sample-efficient representation, where each token is encouraged to capture an independent factor of variation, paving the way for more robust digital biomarkers. | |
| SPLICE: Latent Diffusion over JEPA Embeddings for Conformal Time-Series Inpainting | 2026-04-30 | ShowGenerative models for time-series imputation achieve strong reconstruction accuracy, yet provide no finite-sample reliability guarantees, a critical limitation in power systems where imputed values inform dispatch and planning. We introduce SPLICE (Self-supervised Predictive Latent Inpainting with Conformal Envelopes), a modular framework coupling latent generative imputation with distribution-free, online-adaptive prediction intervals. A JEPA encoder maps daily load segments into a 64-dimensional latent space; a conditional latent bridge with four sampling modes generates candidate gap trajectories; an hourly-conditioned decoder maps back to signal space; and Adaptive Conformal Inference (ACI) wraps the output with coverage-guaranteed prediction bands. The flow-matching variant achieves comparable quality to DDIM in 5--10 ODE steps (5-10x speedup). On thirteen load datasets (nine proprietary, three UCI Electricity, ETTh1), SPLICE achieves the lowest mean Load-only MSE (0.056), winning 9/12 non-degenerate datasets at 91-day gaps and 18/32 across all gap lengths vs. five established baselines, and produces the best CRPS (0.161, -18.3% vs. the strongest competitor). ACI delivers 93--95% empirical coverage, correcting under-coverage failures of up to 7.5 pp observed with static conformal prediction. A pooled JEPA encoder trained on nine feeds transfers to four unseen domains, matching or exceeding per-dataset oracles with only a quick bridge fine-tuning. | |
| Foreclassing: A new machine learning perspective on human decision making with temporal data | 2026-04-30 | ShowTime series forecasts are widely used to inform decisions. Human decision-makers interpret these forecasts, incorporate prior experience and uncertainty about future outcomes, and then make a decision. In this paper, we propose a new machine learning problem, which we call Foreclassing, which addresses settings in which the aim is to automate human involvement in such decision-making processes. Our aim is to develop a unified end-to-end model that takes a time series as input, produces a forecast, accounts for its predictive uncertainty, and makes a downstream classification decision, enabling models to support or automate such temporal decision-making tasks. Related problems arise across a range of applications, yet the literature lacks both a unified methodology and a formal problem statement. By formalizing the task, we aim to stimulate research on such models and encourage cross-domain collaboration. To solve the Foreclassing problem, we propose a deep Bayesian neural network, ForeClassNet. As part of this framework, we introduce a new type of neural network layer, Boltzmann convolutions, which enable probabilistic learning of kernel sizes in convolutional layers. We evaluate the Foreclassing framework against standard time series classification methods and demonstrate the efficacy of ForeClassNet on real-world Foreclassing datasets from the weather, energy, and finance domains, achieving superior performance relative to state-of-the-art time series classifiers. | 20 pa...20 pages, 1 figure, 15 tables |
| Sequential Inference for Gaussian Processes: A Signal Processing Perspective | 2026-04-30 | ShowThe proliferation of capable and efficient machine learning (ML) models marks one of the strongest methodological shifts in signal processing (SP) in its nearly 100-year history. ML models support the development of SP systems that represent complex, nonlinear relationships with high predictive accuracy. Adapting these models often requires sequential inference, which differs both theoretically and methodologically from the usual paradigm of ML, where data are often assumed independent and identically distributed. Gaussian processes (GPs) are a flexible yet principled framework for modeling random functions, and they have become increasingly relevant to SP as statistical and ML methods assume a more prominent role. We provide a self-contained, tutorial-style overview of GPs, with a particular focus on recent methodological advances in sequential, incremental, or streaming inference. We introduce these techniques from a signal-processing perspective while bridging them to recent advances in ML. Many of the developments we survey have direct applications to state-space modeling, sequential regression and forecasting, anomaly detection in time series, sequential Bayesian optimization, adaptive and active sensing, and sequential detection and decision-making. By organizing these advances from a signal-processing perspective, we intend to equip practitioners with practical tools and a coherent roadmap for deploying sequential GP models in real-world systems. | 53 pa...53 pages, 7 figures. Accepted to IEEE Signal Processing Magazine |
| Explainable Load Forecasting with Covariate-Informed Time Series Foundation Models | 2026-04-30 | ShowTime Series Foundation Models (TSFMs) have recently emerged as general-purpose forecasting models and show considerable potential for applications in energy systems. However, applications in critical infrastructure like power grids require transparency to ensure trust and reliability and cannot rely on pure black-box models. To enhance the transparency of TSFMs, we propose an efficient algorithm for computing Shapley Additive Explanations (SHAP) tailored to these models. The proposed approach leverages the flexibility of TSFMs with respect to input context length and provided covariates. This property enables efficient temporal and covariate masking (selectively withholding inputs), allowing for a scalable explanation of model predictions using SHAP. We evaluate two TSFMs - Chronos-2 and TabPFN-TS - on a day-ahead load forecasting task for a transmission system operator (TSO). In a zero-shot setting, both models achieve predictive performance competitive with a Transformer model trained specifically on multiple years of TSO data. The explanations obtained through our proposed approach align with established domain knowledge, particularly as the TSFMs appropriately use weather and calendar information for load prediction. Overall, we demonstrate that TSFMs can serve as transparent and reliable tools for operational energy forecasting. | |
| Early Detection of Water Stress by Plant Electrophysiology: Machine Learning for Irrigation Management | 2026-04-30 | ShowPurpose: Fast detection of plant stress is key to plant phenotyping, precision agriculture, and automated crop management. In particular, efficient irrigation management requires early identification of water stress to optimize resource use while maintaining crop performance. Direct physiological sensing offers the potential to detect stress responses before visible symptoms appear. Methods: In this study, we recorded electrophysiological signals from greenhouse-grown tomato plants subjected to water stress and developed a framework based on machine learning for online stress detection. The recorded time-series data were processed using a processing pipeline that includes statistical feature extraction and selection, automated machine learning or alternatively deep learning, and probability calibration. Results: Across multiple input time horizons, we found that a 30-minute look-back window strikes the best balance between rapid decision-making and classification performance. Using automated machine learning, the framework achieved classification accuracies of up to 92%, outperforming deep learning approaches. Sequential backward selection reduced the feature set while maintaining performance. Importantly, the framework detects transitions from healthy to stressed states in recordings that were not included in the training set. Conclusion: Overall, we provide a decision-support tool for farmers and establish a foundation for biofeedback-driven irrigation control to improve resource efficiency in (semi-)autonomous crop production systems. | |
| ITS-Mina: A Harris Hawks Optimization-Based All-MLP Framework with Iterative Refinement and External Attention for Multivariate Time Series Forecasting | 2026-04-30 | ShowMultivariate time series forecasting plays a pivotal role in numerous real-world applications, including financial analysis, energy management, and traffic planning. While Transformer-based architectures have gained popularity for this task, recent studies reveal that simpler MLP-based models can achieve competitive or superior performance with significantly reduced computational cost. In this paper, we propose ITS-Mina, a novel all-MLP framework for multivariate time series forecasting that integrates three key innovations: (1) an iterative refinement mechanism that progressively enhances temporal representations by repeatedly applying a shared-parameter residual mixer stack, effectively deepening the model's computational capacity without multiplying the number of distinct parameters; (2) an external attention module that replaces traditional self-attention with learnable memory units, capturing cross-sample global dependencies at linear computational complexity; and (3) a Harris Hawks Optimization (HHO) algorithm for automatic dropout rate tuning, enabling adaptive regularization tailored to each dataset. Extensive experiments on six widely-used benchmark datasets demonstrate that ITS-Mina achieves state-of-the-art or highly competitive performance compared to eleven baseline models across multiple forecasting horizons. | 19 pa...19 pages, 2 figures, 3 tables, 4 algorithms |
| Differentiable latent structure discovery for interpretable forecasting in clinical time series | 2026-04-30 | ShowBackground: Timely, uncertainty-aware forecasting from irregular electronic health records (EHR) can support critical-care decisions, yet most approaches either impute to a grid or sacrifice interpretability. We introduce StructGP, a continuous-time multi-task Gaussian process that couples process convolutions with differentiable structure learning to uncover a sparse, ordered directed acyclic graph (DAG) of inter-variable dependencies while preserving principled uncertainty. We further propose LP-StructGP, which augments StructGP with latent pathways-shared, temporally shifted trajectories inferred via subject-specific coupling filters and a softmax gating mechanism-to capture cross-patient progression patterns. Both models are trained under sparsity and acyclicity constraints (augmented Lagrangian, Adam) using scalable low-rank updates. Results: In simulations, the approach reliably recovers ground-truth graphs (Structural Hamming Distance approaching 0 as cohorts grow) and pathway assignments (high Adjusted Rand Index). On a MIMIC-IV septic shock cohort (n=1,008; norepinephrine, creatinine, mean arterial pressure), StructGP improves short-horizon (6 h) forecasting over independent-task baselines (average RMSE 0.68 [95%CI: 0.63--0.74] vs. 0.88 [0.83-0.94]) and, with 15 additional inputs, markedly outperforms unstructured kernels (0.63 [0.58-0.69] vs. 3.02 [2.85-3.18]) with superior calibration (coverage 0.96 vs. 0.84). On the PhysioNet Challenge (12k patients, 41 variables), StructGP attains competitive accuracy (MAE 3.72e-2) relative to a state-of-the-art graph neural model while maintaining calibrated uncertainty. Conclusion: These results show that structured process convolutions with latent pathways deliver interpretable, scalable, and well-calibrated forecasting for irregular clinical time series. | This ...This manuscript is under review at BioData Mining |
| Forecasting the Maintained Score from the OpenSSF Scorecard: A Study of GitHub Repositories Linked to PyPI Packages | 2026-04-30 | ShowBackground: The OpenSSF Scorecard is widely used to assess the security posture of open-source software repositories, with the Maintained metric serving as a key indicator of recent maintenance activities, helping users identify actively maintained projects and potentially abandoned dependencies. However, the metric is inherently retrospective, providing only a short-term snapshot based on the past 90 days of repository activity and offering no insight into the future. This limitation complicates risk assessment for developers and organizations that rely on open-source dependencies. Aims: In this paper, we investigate the feasibility of forecasting future maintenance activities as captured by the OpenSSF Maintained score. Method: Focusing on 3,220 GitHub repositories linked to one of the top 1% most central PyPI libraries, as ranked by PageRank, we reconstruct historical Maintained scores over a three-year period and frame the problem as a multivariate time series forecasting task. We study four target representations: the raw Maintained score (0-10), a bucketed score capturing low (0-2), moderate (3-7), and high (8-10) maintenance levels, the numerical trend slope between consecutive scores, and categorical trend types (downward, stable, upward). We compare a machine learning model (Random Forest) and a deep learning model (LSTM) using training windows of 3-12 months and forecasting horizons of 1-6 months. Results: Our results show that future maintenance activity can be forecasted with meaningful accuracy, particularly when using aggregated representations such as bucketed scores and trend types leading to accuracies above 0.95 and 0.79. Notably, simpler machine learning models perform at least on par with deep learning approaches, suggesting that effective forecasting does not require complex architectures. | 20 pa...20 pages, 8 figures, 1 table |
| Probabilistic Circuits for Irregular Multivariate Time Series Forecasting | 2026-04-30 | ShowJoint probabilistic modeling is essential for forecasting irregular multivariate time series (IMTS) to accurately quantify uncertainty. Existing approaches often struggle to balance model expressivity with consistent marginalization, frequently leading to unreliable or contradictory forecasts. To address this, we propose CircuITS, a novel architecture for probabilistic IMTS forecasting based on probabilistic circuits. Our model is flexible in capturing intricate dependencies between time series channels while structurally guaranteeing valid joint distributions. Experiments on four real world datasets demonstrate that CircuITS achieves superior joint and marginal density estimation compared to state of the art baselines. | |
| Soft-MSM: Differentiable Context-Aware Elastic Alignment for Time Series | 2026-04-30 | ShowElastic distances like dynamic time warping (DTW) are central to time series machine learning because they compare sequences under local temporal misalignment. Soft-DTW is an adaptation of DTW that can be used as a gradient-based loss by replacing the hard minimum in its dynamic-programming recursion with a smooth relaxation. However, this approach does not directly extend to elastic distances whose transition costs depend on the local alignment context. Move-Split-Merge (MSM) is one such distance: it uses context-aware split and merge penalties and has often outperformed DTW in supervised and unsupervised time series machine learning tasks such as classification and clustering. We introduce Soft-MSM, a smooth relaxation of MSM and an elastic alignment loss with context-aware transition costs. Central to the formulation is a smooth gated surrogate for MSM's piecewise split/merge cost, which enables gradients through both the dynamic-programming recursion and the local transition structure. We derive the forward recursion, backward recursion, soft alignment matrix, closed-form gradient, limiting behaviour, and divergence-corrected formulation. Experiments on 112 UCR datasets show that Soft-MSM gives lower MSM barycentre loss than existing MSM barycentre methods, and yields significantly better clustering and nearest-centroid classification performance than Soft-DTW-based alternatives. An implementation is available in the open-source \texttt{aeon} toolkit. | |
| FoReco and FoRecoML: A Unified Toolbox for Forecast Reconciliation in R | 2026-04-30 | ShowForecast reconciliation has become key to improving the accuracy and coherence of forecasts for linearly constrained multiple time series, such as hierarchical and grouped series. Yet, comprehensive software that jointly covers cross-sectional, temporal, and cross-temporal reconciliation has so far been lacking. The R packages FoReco and FoRecoML address this gap by offering a comprehensive and unified framework. The packages respectively implement classical and regression-based linear reconciliation approaches, and non-linear approaches based on machine learning for cross-sectional, temporal and cross-temporal frameworks. Designed for accessibility and flexibility, these packages provide sensible default options that allow new users to apply reconciliation methods with minimal effort, while still giving expert users full control to explore state-of-the-art extensions through customized settings. With this dual focus, FoReco and FoRecoML are versatile tools for practitioners and researchers working on forecast reconciliation. | |
| CausalCompass: Evaluating the Robustness of Time-Series Causal Discovery in Misspecified Scenarios | 2026-04-30 | ShowCausal discovery from time series is a fundamental task in machine learning. However, its widespread adoption is hindered by a reliance on untestable causal assumptions and by the lack of robustness-oriented evaluation in existing benchmarks. To address these challenges, we propose CausalCompass, a flexible and extensible benchmark framework designed to assess the robustness of time-series causal discovery (TSCD) methods under violations of modeling assumptions. To demonstrate the practical utility of CausalCompass, we conduct extensive benchmarking of representative TSCD algorithms across eight assumption-violation scenarios. Our experimental results indicate that no single method consistently attains optimal performance across all settings. Nevertheless, the methods exhibiting superior overall performance across diverse scenarios are almost invariably deep learning-based approaches. We further provide hyperparameter sensitivity analyses to deepen the understanding of these findings. We additionally conduct ablation experiments to explain the strong performance of deep learning-based methods under assumption violations. We also find, somewhat surprisingly, that NTS-NOTEARS relies heavily on standardized preprocessing in practice, performing poorly in the vanilla setting but exhibiting strong performance after standardization. Finally, our work aims to provide a comprehensive and systematic evaluation of TSCD methods under assumption violations, thereby facilitating their broader adoption in real-world applications. The user-friendly implementation, documentation and datasets are available at https://anonymous.4open.science/r/CausalCompass-anonymous-5B4F/. | Major...Major revision from the previous version |
| Gender Bias in YouTube Exposure: Allocative and Structural Inequalities in Political Information Environments | 2026-04-30 | ShowRecommendation algorithms have become the dominant mechanism for information distribution on digital platforms, profoundly shaping personalized information consumption environments. However, gender bias, as a significant form of algorithmic discrimination, may cause users to experience unequal exposure within different political information environments. Taking YouTube as a case, we conduct a controlled social-bot field experiment, where male-coded and female-coded profiles are constructed. We track the exposure and click patterns of these bots to analyze their recommendation trajectories. We analyze the distribution of recommended content from two dimensions: allocative bias and structural bias. First, we find statistically significant differences in allocative bias across male-coded and female-coded profiles, particularly in terms of issue distribution, ideological orientation, and political entities. Secondly, we observe structural bias in the political information environments, characterized by distinct clustering patterns. Additionally, time-series analysis shows that exposure pathways continue to be shaped over time by both communities detected in the co-occurrence network and individual profile-level dynamics. Finally, we construct a simple collaborative-filtering model that reproduces the observed gender bias. We argue that gender bias in recommendation systems is reflected not only in the allocation of political content, but also in how community structures shape these environments, reinforcing societal inequalities and highlighting the need for algorithmic fairness. | 16 pa...16 pages, 6 figures, submit to Information Processing & Management |
| A Study on the Performance of Distributed Training of Data-driven CFD Simulations | 2026-04-30 | ShowData-driven methods for computer simulations are blooming in many scientific areas. The traditional approach to simulating physical behaviors relies on solving partial differential equations (PDE). Since calculating these iterative equations is highly both computationally demanding and time-consuming, data-driven methods leverage artificial intelligence (AI) techniques to alleviate that workload. Data-driven methods have to be trained in advance to provide their subsequent fast predictions, however, the cost of the training stage is non-negligible. This paper presents a predictive model for inferencing future states of a specific fluid simulation that serves as a use case for evaluating different training alternatives. Particularly, this study compares the performance of only CPU, multiGPU, and distributed approaches for training a time series forecasting deep learning (DL) model. With some slight code adaptations, results show and compare, in different implementations, the benefits of distributed GPU-enabled training for predicting high-accuracy states in a fraction of the time needed by the computational fluid dynamics (CFD) solver. | |
| Bayesian inference for hidden Markov models under genuine multimodality with application to ecological time series | 2026-04-30 | ShowBayesian inference in hidden Markov models (HMMs) can be challenging due to the presence of multimodality in the likelihood function, and consequently in the joint posterior distribution, even after correcting for label switching. The parallel tempering (PT) algorithm, a state-space augmentation method, is a widely used approach for dealing with multimodal distributions. Nevertheless, standard implementation of the PT algorithm may not always be sufficient to effectively explore the high-dimensional, complex multimodal posterior distributions that arise in HMMs. In this work, we demonstrate common pitfalls when implementing the PT algorithm for HMMs, approaches to remedy them, and introduce new non-informative prior distributions that facilitate effective posterior distribution exploration. We analyse time series of blue whale dive data with two 3-state HMMs in a Bayesian framework, one of which includes a categorical covariate in the transition probability matrix to account for the effect of sound stimuli on the whale's behavior. We demonstrate how effective implementation of the modified PT algorithm for Bayesian inference leads to effective exploration of the resultant multimodal posterior distribution and how that affects inference for the underlying movement patterns of the blue whales. | 37 pa...37 pages, 11 figures, to be submitted to Bayesian Analysis, corrected author affiliations |
| VTBench: A Multimodal Framework for Time-Series Classification with Chart-Based Representations | 2026-04-29 | ShowTime-series classification (TSC) has advanced significantly with deep learning, yet most models rely solely on raw numerical inputs, overlooking alternative representations. While texture-based encodings such as Gramian Angular Fields (GAF) and Recurrence Plots (RP) convert time series into 2D images, they often require heavy preprocessing and yield less intuitive representations. In contrast, chart-based visualizations offer more interpretable alternatives and show promise in specific domains; however, their effectiveness remains underexplored, with limited systematic evaluation across chart types, visual encoding choices, and datasets. In this work, we introduce VTBench, a systematic and extensible framework that re-examines TSC through multimodal fusion of raw sequences and chart-based visualizations. VTBench generates lightweight, human-interpretable plots -- line, area, bar, and scatter, providing complementary views of the same signal. We develop a modular architecture supporting multiple fusion strategies, including single-chart visual-numerical fusion, multi-chart visual fusion, and full multimodal fusion with raw inputs. Through experiments across 31 UCR datasets, we show that: (1) chart-only models are competitive in selected settings, particularly on smaller datasets; (2) combining multiple chart types can improve accuracy by capturing complementary visual cues; and (3) multimodal models improve or maintain performance when visual features provide non-redundant information, but may degrade accuracy when they introduce redundancy. We further distill practical guidelines for selecting chart types, fusion strategies, and configurations. VTBench establishes a unified foundation for interpretable and effective multimodal time-series classification. | 8 pages main text |
| Preserving Temporal Dynamics in Time Series Generation | 2026-04-29 | ShowTime-series data augmentation plays a crucial role in regression-oriented forecasting tasks, where limited data restricts the performance of deep learning models. While Generative Adversarial Networks (GANs) have shown promise in synthetic time-series generation, existing approaches primarily focus on matching marginal data distributions and often overlook the temporal dynamics that naturally exist in the original multivariate time series. When generating multivariate time series, this mismatch leads to distribution shift and temporal drift, thereby degrading the fidelity of the synthetic sequences. In this work, we propose a model-agnostic Markov Chain Monte Carlo (MCMC)-based framework to mitigate distribution shift and preserve temporal dynamics in synthetic time series. We provide a theoretical analysis of how conditional generative models accumulate deviations under sequential generation and demonstrate that the MCMC algorithm can correct these discrepancies by enforcing consistency with empirical transition statistics between neighboring time points. Extensive experiments on the Lorenz, Licor, ETTh, and ILI datasets using RCGAN, GCWGAN, TimeGAN, SigCWGAN, and AECGAN demonstrate that the proposed MCMC framework consistently improves autocorrelation alignment, skewness error, kurtosis error, R$^2$, discriminative score, and predictive score. These results suggest that synthetic time series consistent with the original data require explicit preservation of transition laws rather than solely relying on adversarial distribution matching, thereby offering a principled direction for improving generative modeling of time-series data. | |
| Context-Aware Graph Attention for Unsupervised Telco Anomaly Detection | 2026-04-29 | ShowWe propose C-MTAD-GAT, an \emph{unsupervised}, \emph{context-aware} graph-attention model for anomaly detection in multivariate time series from mobile networks. C-MTAD-GAT combines graph attention with lightweight context embeddings, and uses a deterministic reconstruction head and multi-step forecaster to produce anomaly scores. Detection thresholds are calibrated \emph{without labels} from validation residuals, keeping the pipeline fully unsupervised. On the public TELCO dataset, C-MTAD-GAT consistently outperforms MTAD-GAT and the Telco-specific DC-VAE, two state-of-the-art baselines, in both event-level and pointwise F1, while triggering substantially fewer alarms. C-MTAD-GAT is also deployed in the Core network of a national mobile operator, demonstrating its resilience in real industrial settings. | |
| LLM-Enhanced Topical Trend Detection at Snapchat | 2026-04-29 | ShowAutomatic detection of topical trends at scale is both challenging and essential for maintaining a dynamic content ecosystem on social media platforms. In this work, we present a large-scale system for identifying emerging topical trends on Snapchat, one of the world's largest short-video social platforms. Our system integrates multimodal topic extraction, time-series burst detection, and LLM-based consolidation and enrichment to enable accurate and timely trend discovery. To the best of our knowledge, this is the first published end-to-end system for topical trend detection on short-video platforms at production scale. Continuous offline human evaluation over six months demonstrates high precision in identifying meaningful trends. The system has been deployed in production at global scale and applied to downstream surfaces including content ranking and search, driving measurable improvements in content freshness and user experience. | |
| Tree-of-Evidence: Efficient "System 2" Search for Faithful Multimodal Grounding | 2026-04-29 | ShowLarge Multimodal Models (LMMs) achieve state-of-the-art performance in high-stakes domains like healthcare, yet their reasoning remains opaque. Current interpretability methods, such as attention mechanisms or post-hoc saliency, often fail to faithfully represent the model's decision-making process, particularly when integrating heterogeneous modalities like time-series and text. We introduce Tree-of-Evidence (ToE), an inference-time search algorithm that frames interpretability as a discrete optimization problem. Rather than relying on soft attention weights, ToE employs lightweight Evidence Bottlenecks that score coarse groups or units of data (e.g., vital-sign windows, report sentences) and performs a beam search to identify the compact evidence set required to reproduce the model's prediction. We evaluate ToE across six tasks spanning three datasets and two domains: four clinical prediction tasks on MIMIC-IV, cross-center validation on eICU, and non-clinical fault detection on LEMMA-RCA. ToE produces auditable evidence traces while maintaining predictive performance, retaining over 0.98 of full-model AUROC with as few as five evidence units across all settings. Under sparse evidence budgets, ToE achieves higher decision agreement and lower probability fidelity error than other approaches. Qualitative analyses show that ToE adapts its search strategy: it often resolves straightforward cases using only vitals, while selectively incorporating text when physiological signals are ambiguous. ToE therefore provides a practical mechanism for auditing multimodal models by revealing which discrete evidence units support each prediction. | |
| IDOBE: Infectious Disease Outbreak forecasting Benchmark Ecosystem | 2026-04-29 | ShowEpidemic forecasting has become an integral part of real-time infectious disease outbreak response. While collaborative ensembles composed of statistical and machine learning models have become the norm for real-time forecasting, standardized benchmark datasets for evaluating such methods are lacking. Further, there is limited understanding on performance of these methods for novel outbreaks with limited historical data. In this paper, we propose IDOBE, a curated collection of epidemiological time series focused on outbreak forecasting. IDOBE compiles from multiple data repositories spanning over a century of surveillance and across U.S. states and global locations. We perform derivative-based segmentation to generate over 10,000 outbreaks covering multiple outcomes such as cases and hospitalizations for 13 diseases. We consider a variety of information-theoretic and distributional measures to quantify the epidemiological diversity of the dataset. Finally, we perform multi-horizon short-term forecasting (1- to 4-week-ahead) through the progression of the outbreak using 11 baseline models and report on their performance. In addition to standard metrics such as NMSE and MAPE for point forecasts, we include probabilistic scoring rules such as Normalized Weighted Interval Score (NWIS) to quantify the performance. We find that MLP-based methods have the most robust performance, with statistical methods having a slight edge during the pre-peak phase. IDOBE dataset along with baselines are released publicly on https://github.com/NSSAC/IDOBE to enable standardized, reproducible benchmarking of outbreak forecasting methods. | 11 pages, 6 figures |
| Exploring the Potential of Probabilistic Transformer for Time Series Modeling: A Report on the ST-PT Framework | 2026-04-29 | ShowThe Probabilistic Transformer (PT) establishes that the Transformer's self-attention plus its feed-forward block is mathematically equivalent to Mean-Field Variational Inference (MFVI) on a Conditional Random Field (CRF). Under this equivalence the Transformer ceases to be a black-box neural network and becomes a programmable factor graph: graph topology, factor potentials, and the message-passing schedule are all explicit and inspectable primitives that can be engineered. PT was originally developed for natural language and in this report we investigate its potential for time series. We first lift PT into the Spatial-Temporal Probabilistic Transformer (ST-PT) to repair PT's missing channel axis and weak per-step semantics, and adopt ST-PT as a shared cornerstone backbone. We then identify three distinct properties that PT/ST-PT offers as a factor-graph model and derive three Research Questions, one per property, that probe how each property can be exploited in time series: RQ1. The graph topology and potentials are direct programmable primitives. Can this be used to inject symbolic time-series priors into ST-PT through structural graph modifications, especially under data scarcity and noise? RQ2. The CRF's factor matrices are the operator's potentials. Can an external condition program these factor matrices on a per-sample basis, so that conditional generation becomes structural rather than feature-level modulation of a fixed one? RQ3. Each MFVI iteration is a Bayesian posterior update on the factor graph. Can this turn the latent transition of latent-space AutoRegressive (AR) forecasting from an opaque MLP into a principled posterior update, and can a CRF teacher distill its latents into the AR student to counter cumulative error? We give one empirical study per question. Together, these three studies position ST-PT as a programmable framework for time-series modeling. | 30 pages, 2 figures |
Trajectory
| Title | Date | Abstract | Comment |
|---|---|---|---|
| OpenSeeker-v2: Pushing the Limits of Search Agents with Informative and High-Difficulty Trajectories | 2026-05-05 | ShowDeep search capabilities have become an indispensable competency for frontier Large Language Model (LLM) agents, yet their development remains dominated by industrial giants. The typical industry recipe involves a highly resource-intensive pipeline spanning pre-training, continual pre-training (CPT), supervised fine-tuning (SFT), and reinforcement learning (RL). In this report, we show that when fueled with informative and high-difficulty trajectories, a simple SFT approach could be surprisingly powerful for training frontier search agents. By introducing three simple data synthesis modifications: scaling knowledge graph size for richer exploration, expanding the tool set size for broader functionality, and strict low-step filtering, we establish a stronger baseline. Trained on merely 10.6k data points, our OpenSeeker-v2 achieves state-of-the-art performance across 4 benchmarks (30B-sized agents with ReAct paradigm): 46.0% on BrowseComp, 58.1% on BrowseComp-ZH, 34.6% on Humanity's Last Exam, and 78.0% on xbench, surpassing even Tongyi DeepResearch trained with heavy CPT+SFT+RL pipeline, which achieves 43.4%, 46.7%, 32.9%, and 75.0%, respectively. Notably, OpenSeeker-v2 represents the first state-of-the-art search agent within its model scale and paradigm to be developed by a purely academic team using only SFT. We are excited to open-source the OpenSeeker-v2 model weights and share our simple yet effective findings to make frontier search agent research more accessible to the community. | 7 pages |
| Evaluating the performance of GCM trajectories using Weather Type frequencies for persistence and transitions: the Iberian Peninsula and Lamb classification | 2026-05-05 | ShowThis study evaluates the performance of 36 historical CMIP6 GCM trajectories (1979-2005) in reproducing atmospheric circulation over the Iberian Peninsula in the summer months (June-September) using the Lamb Weather Type (WT) classification scheme. Using ERA5 reanalysis as the observational reference, we introduce a methodological framework-applicable to any region worldwide-to evaluate GCM performance. This approach extends traditional daily frequency analysis by evaluating both the daily frequency distribution of WTs and their 24-hour dynamic evolution (i.e., transition probabilities and persistence). Model performance is quantified using the Overlap coefficient. A filtering process is applied where only trajectories that successfully reproduce both daily and conditional distributions with a minimum Overlap threshold $t_{sim}$ across a set number of grid points are retained. The findings show that while several models can adequately reproduce daily WT frequencies (16 out of 36), some struggle to capture day-to-day atmospheric transitions. This leads to a final selection of 12 trajectories over the Iberian Peninsula. Model performance across the region is then evaluated using integrated metrics assessing daily reproduction, conditional reproduction, and transition dynamics. Overall, models from the ec earth3 family-specifically the ec earth3 aerchem trajectory-exhibit the best and most consistent performance across the region. Additionally, the results highlight a geographical performance gap: while models generally represent circulation well in the northwest, they face significant challenges in the central and southern Mediterranean regions of the Peninsula. Ultimately, this study establishes that assessing WT persistence and transitions provides a far more discriminative, objective tool for GCM selection than evaluating daily distributions alone. | |
| Harnessing Reasoning Trajectories for Hallucination Detection via Answer-agreement Representation Shaping | 2026-05-05 | ShowLarge reasoning models (LRMs) often generate long, seemingly coherent reasoning traces yet still produce incorrect answers, making hallucination detection challenging. Although trajectories contain useful signals, directly using trace text or vanilla hidden states for detection is brittle: traces vary in form and detectors can overfit to superficial patterns rather than answer validity. We introduce Answer-agreement Representation Shaping (ARS), which learns detection-friendly trace-conditioned representations by explicitly encoding answer stability. ARS generates counterfactual answers through small latent interventions, specifically, perturbing the trace-boundary embedding, and labels each perturbation by whether the resulting answer agrees with the original. It then learns representations that bring answer-agreeing states together and separate answer-disagreeing ones, exposing latent instability indicative of hallucination risk. The shaped embeddings are plug-and-play with existing embedding-based detectors and require no human annotations during training. Experiments demonstrate that ARS consistently improves detection and achieves substantial gains over strong baselines. Code is available at: https://github.com/radiolab-ntu/ars_icml2026. | ICML 2026 |
| TACO: Trajectory Aligning Cross-view Optimisation | 2026-05-05 | ShowCross-View Geo-localisation (CVGL) matches ground imagery against satellite tiles to give absolute position fixes, an alternative to GNSS where signals are occluded, jammed, or spoofed. Recent fine-grained CVGL methods regress sub-tile metric pose, but have only been evaluated as one-shot localisers, never as the primary fix in a live pipeline. Inertial sensing provides high-rate relative motion, but accumulates unbounded drift without an absolute anchor. We propose TACO, a tightly-coupled IMU + fine-grained CVGL pipeline that consumes a single GNSS reading at start-up and thereafter operates on onboard sensing alone. A closed-form cross-track error model triggers CVGL before IMU drift exceeds the matcher's capture radius, and a forward-biased five-point multi-crop search keeps inference cost fixed at five forward passes per fix. A yaw-residual gate rejects fixes that disagree with the onboard compass, and an anisotropic body-frame noise model scales each Unscented Kalman Filter update by per-fix confidence. A factor graph with vetted loop closures provides an offline smoothed trajectory. On the KITTI raw dataset, TACO reduces median Absolute Trajectory Error (ATE) from 97.0m (IMU-only) to 16.3m, a 5.9 times reduction, at <0.1 ms per-frame fusion cost and a 5-10% camera duty cycle. Code is available: github.com/tavisshore/TACO. | |
| Transformer-Guided Deep Reinforcement Learning for Optimal Takeoff Trajectory Design of an eVTOL Drone | 2026-05-04 | ShowThe rapid advancement of electric vertical takeoff and landing (eVTOL) aircraft offers a promising opportunity to alleviate urban traffic congestion but is still limited by excessive power demands, especially during the takeoff phase. Thus, developing optimal takeoff trajectories for minimum energy consumption becomes essential for broader eVTOL aircraft applications. Conventional optimal control methods (such as dynamic programming and linear quadratic regulator) provide highly efficient and well-established solutions but are prohibited by problem dimensionality and complexity. Deep reinforcement learning (DRL) emerges as a special type of artificial intelligence tackling complex, nonlinear systems; however, the training difficulty is a key bottleneck that hinders DRL applications. To address these challenges, we propose the transformer-guided DRL to alleviate the training difficulty by exploring a realistic state space at each time step using a transformer. The proposed transformer-guided DRL was demonstrated on an optimal takeoff trajectory design of an eVTOL drone for minimal energy consumption while meeting takeoff conditions (i.e., minimum vertical displacement and minimum horizontal velocity) by varying control variables (i.e., power and wing angle to the vertical). Results presented that the transformer-guided DRL agent learned to take off with $4.57\times10^6$ time steps, representing $25%$ of the $19.79\times10^6$ time steps needed by a vanilla DRL agent. In addition, the transformer-guided DRL achieved $97.2%$ accuracy on the optimal energy consumption compared against the simulation-based optimal reference, while the vanilla DRL achieved $96.1%$ accuracy. Therefore, the proposed transformer-guided DRL outperformed vanilla DRL in terms of both training efficiency and optimal design verification. | Confe...Conference version with 12 pages and 3 figures |
| OmniTrack++: Omnidirectional Multi-Object Tracking by Learning Large-FoV Trajectory Feedback | 2026-05-04 | ShowTo address panoramic distortion, large search space, and identity ambiguity under a 360° FoV, OmniTrack++ adopts a feedback-driven framework that progressively refines perception with trajectory cues. A DynamicSSM block first stabilizes panoramic features, implicitly alleviating geometric distortion. On top of normalized representations, FlexiTrack Instances use trajectory-informed feedback for flexible localization and reliable short-term association. To ensure long-term robustness, an ExpertTrack Memory consolidates appearance cues via a Mixture-of-Experts design, enabling recovery from fragmented tracks and reducing identity drift. Finally, a Tracklet Management module adaptively switches between end-to-end and tracking-by-detection modes according to scene dynamics, offering a balanced and scalable solution for panoramic MOT. To support rigorous evaluation, we establish the EmboTrack benchmark, a comprehensive dataset for panoramic MOT that includes QuadTrack, captured with a quadruped robot, and BipTrack, collected with a bipedal wheel-legged robot. Together, these datasets span wide-angle environments and diverse motion patterns, providing a challenging testbed for real-world panoramic perception. Extensive experiments on JRDB and EmboTrack demonstrate that OmniTrack++ achieves state-of-the-art performance, yielding substantial HOTA improvements of +3.94 on JRDB and +15.03 on QuadTrack over the original OmniTrack. These results highlight the effectiveness of trajectory-informed feedback, adaptive paradigm switching, and robust long-term memory in advancing panoramic multi-object tracking. Datasets and code will be made available at https://github.com/xifen523/OmniTrack. | Exten...Extended version of CVPR 2025 paper arXiv:2503.04565. Datasets and code will be made publicly available at https://github.com/xifen523/OmniTrack |
| Watermarking LLM Agent Trajectories | 2026-05-04 | ShowLLM agents rely heavily on high-quality trajectory data to guide their problem-solving behaviors, yet producing such data requires substantial task design, high-capacity model generation, and manual filtering. Despite the high cost of creating these datasets, existing literature has overlooked copyright protection for LLM agent trajectories. This gap leaves creators vulnerable to data theft and makes it difficult to trace misuse or enforce ownership rights. This paper introduces ActHook, the first watermarking method tailored for agent trajectory datasets. Inspired by hook mechanisms in software engineering, ActHook embeds hook actions that are activated by a secret input key and do not alter the original task outcome. Like software execution, LLM agents operate sequentially, allowing hook actions to be inserted at decision points without disrupting task flow. When the activation key is present, an LLM agent trained on watermarked trajectories can produce these hook actions at a significantly higher rate, enabling reliable black-box detection. Experiments on mathematical reasoning, web searching, and software engineering agents show that ActHook achieves an average detection AUC of 94.3 on Qwen-2.5-Coder-7B while incurring negligible performance degradation. | 23 pa...23 pages, 10 figures; accepted by ICML 2026 |
| Statistical description and dimension reduction of continuous time categorical trajectories with multivariate functional principal components | 2026-05-03 | ShowGetting tools that allow simple representations and comparisons of a set of categorical trajectories is of major interest for statisticians. Without loosing any information, we associate to each state a binary random indicator function, taking values in ${0,1}$, and turn the problem of statistical description of the categorical trajectories into a multivariate functional principal components analysis. This viewpoint encompasses experimental frameworks where two or more states can be observed simultaneously. The sample paths being piecewise constant, with a finite number of jumps, this a rare case in functional data analysis in which the trajectories are not supposed to be continuous and can be observed exhaustively. Under the weak hypothesis assuming only continuity in probability of the $0-1$ trajectories, the means and the (multivariate) covariance function are continuous and have interpretations in terms of departure from independence of the joint probabilities. Considering a functional data point of view, we show that the binary trajectories, which are right-continuous functions with left-hand limits, can be seen as random elements in the Hilbert space of square integrable functions. The multivariate functional principal components are simple to interpret and we show that we can get consistent estimators of the mean trajectories and the covariance functions under weak regularity assumptions. The ability of the approach to represent categorical trajectories in a small dimension space is illustrated on a data set of sensory perceptions, considering different gustometer-controlled stimuli experiments. | |
| Optimizing Trajectory-Trees in Belief Space: An Application from Model Predictive Control to Task and Motion Planning | 2026-05-03 | ShowThis paper explores the benefits of computing arborescent trajectories (trajectory-trees) instead of commonly used sequential trajectories for partially observable robotic planning problems. In such environments, a robot infers knowledge from observations, and the optimal course of action depends on these observations. \revise{Trajectory-trees, optimized in belief space, naturally capture this dependency by branching where the belief state is expected to evolve into multiple distinct scenarios, such as upon receiving an observation. Unlike sequential trajectories, which model a single forward evolution of the system, trajectory-trees capture multiple possible contingencies.} First, we focus on Model Predictive Control (MPC) and demonstrate the benefits of planning tree-like trajectories. We formulate the control problem as the optimization of a tree with a single branching (PO-MPC). This improves performance by reducing control costs through more informed planning. To satisfy the real-time constraints of MPC, we develop an optimization algorithm called Distributed Augmented Lagrangian (D-AuLa), which leverages the decomposability of the PO-MPC formulation to parallelize and accelerate the optimization. We apply the method to both linear and non-linear MPC problems using autonomous driving examples. Second, we address Task And Motion Planning (TAMP), and introduce a planner (PO-LGP) reasoning on decision trees at task level, and trajectory-trees at motion-planning level. This approach builds upon the Logic-Geometric-Programming Framework (LGP) and extends it to partially observable problems. The experiments show the method's applicability to problems with a small belief state size, and scales to larger problems by optimizing explorative policies, which are used as macro-actions in an overarching task plan. | 41 pages |
| LeapAlign: Post-Training Flow Matching Models at Any Generation Step by Building Two-Step Trajectories | 2026-05-03 | ShowThis paper focuses on the alignment of flow matching models with human preferences. A promising way is fine-tuning by directly backpropagating reward gradients through the differentiable generation process of flow matching. However, backpropagating through long trajectories results in prohibitive memory costs and gradient explosion. Therefore, direct-gradient methods struggle to update early generation steps, which are crucial for determining the global structure of the final image. To address this issue, we introduce LeapAlign, a fine-tuning method that reduces computational cost and enables direct gradient propagation from reward to early generation steps. Specifically, we shorten the long trajectory into only two steps by designing two consecutive leaps, each skipping multiple ODE sampling steps and predicting future latents in a single step. By randomizing the start and end timesteps of the leaps, LeapAlign leads to efficient and stable model updates at any generation step. To better use such shortened trajectories, we assign higher training weights to those that are more consistent with the long generation path. To further enhance gradient stability, we reduce the weights of gradient terms with large magnitude, instead of completely removing them as done in previous works. When fine-tuning the Flux model, LeapAlign consistently outperforms state-of-the-art GRPO-based and direct-gradient methods across various metrics, achieving superior image quality and image-text alignment. | Accep...Accepted by CVPR 2026. Project page: https://rockeycoss.github.io/leapalign/ |
| TrajShield: Trajectory-Level Safety Mediation for Defending Text-to-Video Models Against Jailbreak Attacks | 2026-05-03 | ShowText-to-Video (T2V) models have demonstrated remarkable capability in generating temporally coherent videos from natural language prompts, yet they also risk producing unsafe content such as violence or explicit material. Existing prompt-level defenses are largely inherited from text-to-image safety and operate on the lexical surface of the input, making them vulnerable to jailbreak attacks that disguise harmful intent through rephrasing or adversarial prompting. Moreover, T2V generation introduces a distinctive challenge overlooked by prior work: temporally emergent risk, where a seemingly benign prompt leads to unsafe content through the generator's temporal extrapolation toward narrative coherence. We propose \method{}, a training-free, inference-time defense framework that reformulates T2V safety as a causal intervention in a temporally structured semantic space. TrajShield handles explicit unsafe prompts, jailbreak attacks, and temporally emergent risks in a unified manner by simulating the implied trajectory of a prompt, localizing the causal origin of potential risk, and applying a minimally invasive rewrite that neutralizes the risk while preserving safety-irrelevant semantics. Experiments on T2VSafetyBench across 14 safety categories and multiple T2V backends demonstrate that TrajShield achieves state-of-the-art defenseive performance while maintaining high semantic fidelity, substantially outperforming existing defenses, with an average ASR reduction of 52.44%. | |
| InsTraj: Instructing Diffusion Models with Travel Intentions to Generate Real-world Trajectories | 2026-05-02 | ShowThe generation of realistic and controllable GPS trajectories is a fundamental task for applications in urban planning, mobility simulation, and privacy-preserving data sharing. However, existing methods face a two-fold challenge: they lack the deep semantic understanding to interpret complex user travel intent, and struggle to handle complex constraints while maintaining the realistic diversity inherent in human behavior. To resolve this, we introduce InsTraj, a novel framework that instructs diffusion models to generate high-fidelity trajectories directly from natural language descriptions. Specifically, InsTraj first utilizes a powerful large language model to decipher unstructured travel intentions formed in natural language, thereby creating rich semantic blueprints and bridging the representation gap between intentions and trajectories. Subsequently, we proposed a multimodal trajectory diffusion transformer that can integrate semantic guidance to generate high-fidelity and instruction-faithful trajectories that adhere to fine-grained user intent. Comprehensive experiments on real-world datasets demonstrate that InsTraj significantly outperforms state-of-the-art methods in generating trajectories that are realistic, diverse, and semantically faithful to the input instructions. | |
| Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbreak Detection | 2026-05-02 | ShowRepresentation Engineering typically relies on static refusal vectors derived from terminal representations. We move beyond this paradigm, demonstrating that refusal is a dynamic and sparse process rather than a localized outcome. Using Causal Tracing, we uncover the Refusal Trajectory-a persistent upstream signature that remains intact even when adversarial attacks (e.g., GCG) suppress terminal signals. Leveraging this, we propose SALO (Sparse Activation Localization Operator), an inference-time detector designed to capture these latent patterns. SALO effectively recovers defense capabilities against forced-decoding attacks, improving detection rates from ~0% to >90% where methods relying on terminal states perform poorly. | Accep...Accepted to the 43rd International Conference on Machine Learning (ICML 2026). Pre-camera-ready version |
| Latent Trajectory Dynamics in Large Language Models: A Manifold Evolution Framework with Empirical Validation | 2026-05-02 | ShowUnderstanding how latent representations evolve during generation is a central open problem in large language model interpretability. We introduce \textbf{Dynamical Manifold Evolution Theory} (DMET), a phenomenological framework that models LLM generation as a controlled dynamical system evolving along a trajectory on a low-dimensional semantic manifold. DMET formalizes the structural correspondence between Transformer components and a first-order ODE governed by a semantic potential $V$, and characterizes trajectory geometry through three falsifiable proxy metrics: state continuity $C$, attractor clustering quality $Q$, and topological persistence $P$, targeting local smoothness, meso-scale basin structure, and global topological organization, respectively. Across six model architectures, four task types, and 1,080 experimental runs, all three metrics consistently predict text quality outcomes -- log-perplexity, grammaticality, and cross-sentence coherence -- after controlling for decoding parameters, with associations surviving Benjamini--Hochberg correction. Ablation and sanity-check experiments confirm that the effects arise from genuine trajectory structure rather than static distributional artefacts. Furthermore, online monitoring of $C$ drives an adaptive decoding controller that reduces perplexity from 48.5 to 14.6 relative to a fixed-parameter baseline, demonstrating that latent dynamics characterization translates directly into actionable generation control. | |
| MTA: Multi-Granular Trajectory Alignment for Large Language Model Distillation | 2026-05-02 | ShowKnowledge distillation is a key technique for compressing large language models (LLMs), but most existing methods align representations at fixed layers or token-level outputs, ignoring how representations evolve across depth. As a result, the student is only weakly guided to capture the teacher's internal relational structure during distillation, which limits knowledge transfer. To address this limitation, we propose Multi-Granular Trajectory Alignment (MTA), a framework that aligns teacher and student representations along their layer-wise transformation trajectory. MTA adopts a layer-adaptive strategy: lower layers are aligned at the word level to preserve lexical information, while higher layers operate on phrase-level spans (e.g., noun and verb phrases) to capture compositional semantics. We instantiate this idea through a Dynamic Structural Alignment loss that matches the relative geometry among semantic units within each layer. This design is motivated by empirical findings that Transformer representations become increasingly abstract with depth, and is also consistent with linguistic views in which higher-level meaning emerges through the composition of lower-level lexical units. We further incorporate a Hidden Representation Alignment loss to directly align selected teacher-student layers. Experiments show that MTA consistently outperforms state-of-the-art baselines on standard benchmarks, with ablations confirming the contribution of each component. | ACL 2026 |
| Uncertainty-Aware Trip Purpose Inference from GPS Trajectories via POI Semantic Zones and Pareto Calibration | 2026-05-02 | ShowLarge-scale GPS trajectory data offer rich observations of human mobility, yet assigning trip purposes to detected stops remains challenging due to the absence of individual-level ground truth, spatial uncertainty from GPS noise and incomplete points of interest (POIs) coverage, and fundamental behavioral differences across trip purposes. We propose a weakly supervised framework integrating neighborhood-level POI semantic zones with distance-weighted spatial likelihoods, differentiated inference strategies for mandatory and non-mandatory activities, and a multi-phase Pareto optimization that jointly minimizes distributional divergence from household travel survey statistics and maximizes inference reliability without requiring annotated labels. Evaluated on over 81 million staypoints in Los Angeles, the framework reduces activity type frequency Jensen-Shannon distance (JSD) by 23%, start time JSD by 48%, and duration JSD by 12% respectively relative to a comparable baseline. The proposed approach provides a scalable and uncertainty-aware path from raw GPS trajectories to semantically annotated mobility data for travel demand modeling and transportation policy analysis. | |
| Deep Kernel Learning for Stratifying Glaucoma Trajectories | 2026-05-01 | ShowEffectively stratifying patient risk in chronic diseases like glaucoma is a major clinical challenge. Clinicians need tools to identify patients at high risk of progression from sparse and irregularly-sampled electronic health records (EHRs). We propose a novel deep kernel learning (DKL) architecture that leverages a Gaussian Process (GP) backend. The GP's kernel is defined by a transformer-based feature extractor applied to clinical-BERT embeddings to model glaucoma patient trajectories from multimodal EHR data. Our method successfully identifies three clinically distinct patient subgroups. Crucially, the model learns to decouple disease progression from current severity, identifying a high-risk group with a worsening trajectory despite having better average visual acuity than a second, stably poor group. This reveals that the model learns to identify progression risk rather than just the current disease state. This ability to stratify patients based on their risk trajectory progression offers a powerful tool for clinical decision support, enabling targeted interventions for high-risk individuals and improving the management of glaucoma care. | |
| GSDrive: Reinforcing Driving Policies by Multi-mode Trajectory Probing with 3D Gaussian Splatting Environment | 2026-05-01 | ShowEnd-to-end (E2E) autonomous driving presents a promising approach for translating perceptual inputs directly into driving actions. However, prohibitive annotation costs and temporal data quality degradation hinder long-term real-world deployment. While combining imitation learning (IL) and reinforcement learning (RL) is a common strategy for policy improvement, conventional RL training relies on delayed, event-based rewards-policies learn only from catastrophic outcomes such as collisions, leading to premature convergence to suboptimal behaviors. To address these limitations, we introduce GSDrive, a framework that exploits 3D Gaussian Splatting (3DGS) for differentiable, physics-based reward shaping in E2E driving policy improvement. Our method incorporates a flow matching-based trajectory predictor within the 3DGS simulator, enabling multi-mode trajectory probing where candidate trajectories are rolled out to assess prospective rewards. This establishes a bidirectional knowledge exchange between IL and RL by grounding reward functions in physically simulated interaction signals, offering immediate dense feedback instead of sparse catastrophic events. Evaluated on the reconstructed nuScenes dataset, our method surpasses existing simulation-based RL driving approaches in closed-loop experiments. Code is available at https://github.com/ZionGo6/GSDrive. | initial version |
| Modeling Clinical Concern Trajectories in Language Model Agents | 2026-04-30 | ShowLarge language model (LLM) agents deployed in clinical settings often exhibit abrupt, threshold-driven behavior, offering little visibility into accumulating risk prior to escalation. In real-world care, however, clinicians act on gradually rising concern rather than instantaneous triggers. We study whether explicit state dynamics can expose such pre-escalation signals without delegating clinical authority to the agent. We introduce a lightweight agent architecture in which a memoryless clinical risk encoder is integrated over time using first- and second-order dynamics to produce a continuous escalation pressure signal. Across synthetic ward scenarios, stateless agents exhibit sharp escalation cliffs, while second-order dynamics produce smooth, anticipatory concern trajectories despite similar escalation timing. These trajectories surface sustained unease prior to escalation, enabling human-in-the-loop monitoring and more informed intervention. Our results suggest that explicit state dynamics can make LLM agents more clinically legible by revealing how long concern has been rising, not just when thresholds are crossed. | |
| Training-Free Reward-Guided Image Editing via Trajectory Optimal Control | 2026-04-30 | ShowRecent advancements in diffusion and flow-matching models have demonstrated remarkable capabilities in high-fidelity image synthesis. A prominent line of research involves reward-guided guidance, which steers the generation process during inference to align with specific objectives. However, leveraging this reward-guided approach to the task of image editing, which requires preserving the semantic content of the source image while enhancing a target reward, is largely unexplored. In this work, we introduce a novel framework for training-free, reward-guided image editing. We formulate the editing process as a trajectory optimal control problem where the reverse process of a diffusion model is treated as a controllable trajectory originating from the source image, and the adjoint states are iteratively updated to steer the editing process. Through extensive experiments across distinct editing tasks, we demonstrate that our approach significantly outperforms existing inversion-based training-free guidance baselines, achieving a superior balance between reward maximization and fidelity to the source image without reward hacking. | Poste...Poster in ICLR 2026; 22 pages, 9 figures. The code is available at https://github.com/jinhojsk515/ITOC |
| Adaptive Nonlinear MPC for Trajectory Tracking of An Overactuated Tiltrotor Hexacopter | 2026-04-30 | ShowOmnidirectional micro aerial vehicles (OMAVs) are more capable of doing environmentally interactive tasks due to their ability to exert full wrenches while maintaining stable poses. However, OMAVs often incorporate additional actuators and complex mechanical structures to achieve omnidirectionality. Obtaining precise mathematical models is difficult, and the mismatch between the model and the real physical system is not trivial. The large model-plant mismatch significantly degrades overall system performance if a non-adaptive model predictive controller (MPC) is used. This work presents the $\mathcal{L}_1$-MPC, an adaptive nonlinear model predictive controller for accurate 6-DOF trajectory tracking of an overactuated tiltrotor hexacopter in the presence of model uncertainties and external disturbances. The $\mathcal{L}_1$-MPC adopts a cascaded system architecture in which a nominal MPC is followed and augmented by an $\mathcal{L}_1$ adaptive controller. The proposed method is evaluated against the non-adaptive MPC, the EKF-MPC, and the PID method in both numerical and PX4 software-in-the-loop simulation with Gazebo. The $\mathcal{L}_1$-MPC reduces the tracking error by around 90% when compared to a non-adaptive MPC, and the $\mathcal{L}_1$-MPC has lower tracking errors, higher uncertainty estimation rates, and less tuning requirements over the EKF-MPC. We will make the implementations, including the hardware-verified PX4 firmware and Gazebo plugins, open-source at https://github.com/HITSZ-NRSL/omniHex. | (1) E...(1) Eq. (10) sign error, inconsistent with Eq. (14). (2) Eq. (15) spurious Coriolis term (skips transport theorem). (3) typo before Eq. (21): _Bω_dot_EKF?_Bτ_dot_EKF. (4) Sec. IV comparison lacks systematic tuning and does not support its claims. (5) the open-source release at github.com/HITSZ-NRSL/omniHex will not happen |
| Frozen LLMs as Map-Aware Spatio-Temporal Reasoners for Vehicle Trajectory Prediction | 2026-04-30 | ShowLarge language models (LLMs) have recently demonstrated strong reasoning capabilities and attracted increasing research attention in the field of autonomous driving (AD). However, safe application of LLMs on AD perception and prediction still requires a thorough understanding of both the dynamic traffic agents and the static road infrastructure. To this end, this study introduces a framework to evaluate the capability of LLMs in understanding the behaviors of dynamic traffic agents and the topology of road networks. The framework leverages frozen LLMs as the reasoning engine, employing a traffic encoder to extract spatial-level scene features from observed trajectories of agents, while a lightweight Convolutional Neural Network (CNN) encodes the local high-definition (HD) maps. To assess the intrinsic reasoning ability of LLMs, the extracted scene features are then transformed into LLM-compatible tokens via a reprogramming adapter. By residing the prediction burden with the LLMs, a simpler linear decoder is applied to output future trajectories. The framework enables a quantitative analysis of the influence of multi-modal information, especially the impact of map semantics on trajectory prediction accuracy, and allows seamless integration of frozen LLMs with minimal adaptation, thereby demonstrating strong generalizability across diverse LLM architectures and providing a unified platform for model evaluation. | Accep...Accepted for publication at IEEE Intelligent Vehicles Symposium 2026 |
| FeaXDrive: Feasibility-aware Trajectory-Centric Diffusion Planning for End-to-End Autonomous Driving | 2026-04-30 | ShowEnd-to-end diffusion planning has shown strong potential for autonomous driving, but the physical feasibility of generated trajectories remains insufficiently addressed. In particular, generated trajectories may exhibit local geometric irregularities, violate trajectory-level kinematic constraints, or deviate from the drivable area, indicating that the commonly used noise-centric formulation in diffusion planning is not yet well aligned with the trajectory space where feasibility is more naturally characterized. To address this issue, we propose FeaXDrive, a feasibility-aware trajectory-centric diffusion planning method for end-to-end autonomous driving. The core idea is to treat the clean trajectory as the unified object for feasibility-aware modeling throughout the diffusion process. Built on this trajectory-centric formulation, FeaXDrive integrates adaptive curvature-constrained training to improve intrinsic geometric and kinematic feasibility, drivable-area guidance within reverse diffusion sampling to enhance consistency with the drivable area, and feasibility-aware GRPO post-training to further improve planning performance while balancing trajectory-space feasibility. Experiments on the NAVSIM benchmark show that FeaXDrive achieves strong closed-loop planning performance while substantially improving trajectory-space feasibility. These findings highlight the importance of explicitly modeling trajectory-space feasibility in end-to-end diffusion planning and provide a step toward more reliable and physically grounded autonomous driving planners. | 22 pages, 6 figures |
| Dynamic-TD3: A Novel Algorithm for UAV Path Planning with Dynamic Obstacle Trajectory Prediction | 2026-04-30 | ShowDeep reinforcement learning (DRL) finds extensive application in autonomous drone navigation within complex, high-risk environments. However, its practical deployment faces a safety-exploration dilemma: soft penalty mechanisms encourage risky trial-and-error, while most constraint-based methods suffer degraded performance under sensor noise and intent uncertainty. We propose Dynamic-TD3, a physically enhanced framework that enforces strict safety constraints while maintaining maneuverability by modeling navigation as a Constrained Markov Decision Process (CMDP). This framework integrates an Adaptive Trajectory Relational Evolution Mechanism (ATREM) to capture long-range intentions and employs a Physically Aware Gated Kalman Filter (PAG-KF) to mitigate non-stationary observation noise. The resulting state representation drives a dual-criterion policy that balances mission efficiency against hard safety constraints via Lagrangian relaxation. In experiments with aggressive dynamic threats, this approach demonstrates superior collision avoidance performance, reduced energy consumption, and smoother flight trajectories. | 6 pages, 5 figures |
| Language-Conditioned Safe Trajectory Generation for Spacecraft Rendezvous | 2026-04-29 | ShowReliable real-time trajectory generation is essential for future autonomous spacecraft. While recent progress in nonconvex guidance and control is paving the way for onboard autonomous trajectory optimization, these methods still rely on extensive expert input (e.g., waypoints, constraints, mission timelines, etc.), which limits operational scalability in complex missions such as rendezvous and proximity operations. This paper introduces SAGES (Semantic Autonomous Guidance Engine for Space), a trajectory-generation framework that translates natural-language commands into spacecraft trajectories that reflect high-level intent while respecting nonconvex constraints. Experiments in two settings (fault-tolerant proximity operations with continuous-time constraint enforcement and a free-flying robotic platform) demonstrate that SAGES reliably produces trajectories aligned with human commands, achieving over 90% semantic-behavioral consistency across diverse behavior modes. Ultimately, this work marks an initial step toward language-conditioned, constraint-aware spacecraft trajectory generation, enabling operators to interactively guide both safety and behavior through intuitive natural-language commands with reduced expert burden. Project Website: https://semantic-guidance4space.github.io/ | 42 pa...42 pages, 12 figures. Submitted to AIAA Journal of Guidance, Control, and Dynamics |
| Global Sampling-Based Trajectory Optimization for Contact-Rich Manipulation via KernelSOS | 2026-04-29 | ShowContact-rich manipulation is challenging due to its high dimensionality, the requirement for long time horizons, and the presence of hybrid contact dynamics. Sampling-based methods have become a popular approach for this class of problems, but without explicit mechanisms for global exploration, they are susceptible to converging to poor local minima. In this paper, we introduce Global-MPPI, a unified trajectory optimization framework that integrates global exploration and local refinement. At the global level, we leverage kernel sum-of-squares optimization to identify globally promising regions of the solution space. To enable reliable performance for the non-smooth landscapes inherent to contact-rich manipulation, we introduce a graduated non-convexity strategy based on log-sum-exp smoothing, which transitions the optimization landscape from a smoothed surrogate to the original non-smooth objective. Finally, we employ the model-predictive path integral method to locally refine the solution. We evaluate Global-MPPI on high-dimensional, long-horizon contact-rich tasks, including the PushT task and dexterous in-hand manipulation. Experimental results demonstrate that our approach robustly uncovers high-quality solutions, achieving faster convergence and lower final costs compared to existing baseline methods. | 8 pages, 5 figures |
| A Dual Perspective on Synthetic Trajectory Generators: Utility Framework and Privacy Vulnerabilities | 2026-04-29 | ShowHuman mobility data are used in numerous applications, ranging from public health to urban planning. Human mobility is inherently sensitive, as it can contain information such as religious beliefs and political affiliations. Historically, it has been proposed to modify the information using techniques such as aggregation, obfuscation, or noise addition, to adequately protect privacy and eliminate concerns. As these methods come at a great cost in utility, new methods leveraging development in generative models, were introduced. The extent to which such methods answer the privacy-utility trade-off remains an open problem. In this paper, we introduced a first step towards solving it, by the introduction and application of a new framework for utility evaluation. Furthermore, we provide evidence that privacy evaluation remains a great challenge to consider and that it should be tackled through adversarial evaluation in accordance with the current EU regulation. We propose a new membership inference attack against a subcategory of generative models, even though this subcategory was deemed private due to its resistance over the trajectory user-linking problem. | |
| Benchmarks for Trajectory Safety Evaluation and Diagnosis in OpenClaw and Codex: ATBench-Claw and ATBench-Codex | 2026-04-29 | ShowAs agent systems move into increasingly diverse execution settings, trajectory-level safety evaluation and diagnosis require benchmarks that evolve with them. ATBench is a diverse and realistic agent trajectory benchmark for safety evaluation and diagnosis. This report presents ATBench-Claw and ATBench-Codex, two domain-customized extensions that carry ATBench into the OpenClaw and OpenAI Codex / Codex-runtime settings. The key adaptation mechanism is to analyze each new setting, customize the three-dimensional Safety Taxonomy over risk source, failure mode, and real-world harm, and then use that customized taxonomy to define the benchmark specification consumed by the shared ATBench construction pipeline. This extensibility matters because agent frameworks remain relatively stable at the architectural level even as their concrete execution settings, tool ecosystems, and product capabilities evolve quickly. Concretely, ATBench-Claw targets OpenClaw-sensitive execution chains over tools, skills, sessions, and external actions, while ATBench-Codex targets trajectories in the OpenAI Codex / Codex-runtime setting over repositories, shells, patches, dependencies, approvals, and runtime policy boundaries. Our emphasis therefore falls on taxonomy customization, domain-specific risk coverage, and benchmark design under a shared ATBench generation framework. | 18 pages, 3 figures |
| Enforcing Benign Trajectories: A Behavioral Firewall for Structured-Workflow AI Agents | 2026-04-29 | ShowStructured-workflow agents driven by large language models execute tool calls against sensitive external environments. We propose \codename, a telemetry-driven behavioral anomaly detection firewall. Drawing on sequence-based intrusion detection, \codename\ compiles verified benign tool-call telemetry into a parameterized deterministic finite automaton (pDFA). The model defines permitted tool sequences, sequential contexts, and parameter bounds. At runtime, a lightweight gateway enforces these boundaries via an $O(1)$ state-transition structural lookup, shifting computationally expensive analysis entirely offline. Evaluated on the Agent Security Bench (ASB), \codename\ achieves a 5.6% macro-averaged attack success rate (ASR) across five scenarios. Within three structured workflows, ASR drops to 2.2%, outperforming Aegis, a state-of-the-art stateless scanner, at 12.8%. \codename\ achieves 0% ASR on multi-step and context-sequential attacks in structured settings. Furthermore, against 1,000 algorithmically spliced exfiltration payloads, only 1.4% matched valid structural paths, all of which failed end-to-end string parameter guards (0 successes out of 14 surviving paths, 95% CI [0%, 23.2%]). \codename\ introduces just 2.2~ms of per-call latency (a 3.7$\times$ speedup over \textsc{Aegis}) while maintaining a 2.0% benign task failure rate (BTFR) on benign workloads. Modeling the behavioral trajectory effectively collapses the available attack surface, but unmaintained continuous parameter bounds remain vulnerable to synonym-substitution attacks (18% evasion rate). Thus, exact-match whitelisting of sensitive parameters ultimately bears the final defensive load against execution. | |
| OnSiteVRU: A High-Resolution Trajectory Dataset for High-Density Vulnerable Road Users | 2026-04-29 | ShowWith the acceleration of urbanization and the growth of transportation demands, the safety of vulnerable road users (VRUs, such as pedestrians and cyclists) in mixed traffic flows has become increasingly prominent, necessitating high-precision and diverse trajectory data to support the development and optimization of autonomous driving systems. However, existing datasets fall short in capturing the diversity and dynamics of VRU behaviors, making it difficult to meet the research demands of complex traffic environments. To address this gap, this study developed the OnSiteVRU datasets, which cover a variety of scenarios, including intersections, road segments, and urban villages. These datasets provide trajectory data for motor vehicles, electric bicycles, and human-powered bicycles, totaling approximately 17,429 trajectories with a precision of 0.04 seconds. The datasets integrate both aerial-view natural driving data and onboard real-time dynamic detection data, along with environmental information such as traffic signals, obstacles, and real-time maps, enabling a comprehensive reconstruction of interaction events. The results demonstrate that VRU_Data outperforms traditional datasets in terms of VRU density and scene coverage, offering a more comprehensive representation of VRU behavioral characteristics. This provides critical support for traffic flow modeling, trajectory prediction, and autonomous driving virtual testing. The dataset is publicly available for download at: https://www.kaggle.com/datasets/zcyan2/mixed-traffic-trajectory-dataset-in-from-shanghai. | |
| From Threads to Trajectories: A Multi-LLM Pipeline for Community Knowledge Extraction from GitHub Issue Discussions | 2026-04-28 | ShowResolution of complex post-production issues in large-scale open-source software (OSS) projects requires significant cognitive effort, as developers need to go through long, unstructured and fragmented issue discussion threads before that. In this paper, we present SWE-MIMIC-Bench, an issue trajectory dataset generated from raw GitHub discussions using an automated multi-LLM pipeline. Unlike simple summarization, this pipeline utilizes a group of closed-source LLMs to perform granular tasks: analyzing individual comments with awareness of externally-linked resources, classifying comment analyses into label-specific fields (e.g., root cause, solution plan, implementation progress), and synthesizing label-aware trajectories which capture a structured and coherent narrative of the entire discussion thread. Our pipeline uses five closed-source LLM configurations for distinct purposes: label classification, inline code block and external link summarization, comment analysis, label-specific field classification and trajectory synthesis. By generating concise and reliable trajectories from complex conversation threads, this system can assist developers and researchers of broader software engineering community to understand the experience-driven collaborative approach for issue diagnosis. Furthermore, the generated trajectories can be used to train modern LLM agents to think and act like an expert developer. We evaluated our system on 800 real-world GitHub issues drawn from the SWE-Bench-Pro, SWE-Bench-Multilingual and SWE-Bench-Verified dataset, achieving a 91.7% success rate in extracting 734 high-fidelity reasoning trajectories. | |
| Quantum Dynamics via Score Matching on Bohmian Trajectories | 2026-04-28 | ShowWe solve the time-dependent Schrödinger equation by learning the score function, the gradient of the log-probability density, on Bohmian trajectories. In Bohm's formulation of quantum mechanics, particles follow deterministic paths under the classical potential supplemented by a quantum potential depending on the score function of the evolving density. These non-crossing Bohmian trajectories form a continuous normalizing flow governed by the score. We parametrize the score with a neural network and minimize a self-consistent Fisher divergence between the network and the score of the resulting density. We prove that the zero-loss minimizer of this self-consistent objective recovers Schrödinger dynamics for nodeless wave functions, a condition naturally met in quantum vibrations of atoms. We demonstrate the approach on wavepacket splitting in a double-well potential and anharmonic vibrations of a Morse chain. By recasting real-time quantum dynamics as a self-consistent score-driven normalizing flow, this framework opens the time-dependent Schrödinger equation to the rapidly advancing toolkit of modern generative modeling. | 8 pag...8 pages, 5 figues, code at https://github.com/wangleiphy/BohmianFlow |
| Sparse Personalized Text Generation with Multi-Trajectory Reasoning | 2026-04-27 | ShowAs Large Language Models (LLMs) advance, personalization has become a key mechanism for tailoring outputs to individual user needs. However, most existing methods rely heavily on dense interaction histories, making them ineffective in cold-start scenarios where such data is sparse or unavailable. While external signals (e.g., content of similar users) can offer a potential remedy, leveraging them effectively remains challenging: raw context is often noisy, and existing methods struggle to reason over heterogeneous data sources. To address these issues, we introduce PAT (Personalization with Aligned Trajectories), a reasoning framework for cold-start LLM personalization. PAT first retrieves information along two complementary trajectories: writing-style cues from stylistically similar users and topic-specific context from preference-aligned users. It then employs a reinforcement learning-based, iterative dual-reasoning mechanism that enables the LLM to jointly refine and integrate these signals. Experimental results across real-world personalization benchmarks show that PAT consistently improves generation quality and alignment under sparse-data conditions, establishing a strong solution to the cold-start personalization problem. | |
| Bayesian Inverse Transition Learning: Learning Dynamics From Near-Optimal Trajectories | 2026-04-27 | ShowWe consider the problem of estimating the transition dynamics $T^$ from near-optimal expert trajectories in the context of offline model-based reinforcement learning. We develop a novel constraint-based method, Inverse Transition Learning, that treats the limited coverage of the expert trajectories as a \emph{feature}: we use the fact that the expert is near-optimal to inform our estimate of $T^$. We integrate our constraints into a Bayesian approach. Across both synthetic environments and real healthcare scenarios like Intensive Care Unit (ICU) patient management in hypotension, we demonstrate not only significant improvements in decision-making, but that our posterior can inform when transfer will be successful. | |
| VERTIGO: Visual Preference Optimization for Cinematic Camera Trajectory Generation | 2026-04-27 | ShowCinematic camera control relies on a tight feedback loop between director and cinematographer, where camera motion and framing are continuously reviewed and refined. Recent generative camera systems can produce diverse, text-conditioned trajectories, but they lack this "director in the loop" and have no explicit supervision of whether a shot is visually desirable. This results in in-distribution camera motion but poor framing, off-screen characters, and undesirable visual aesthetics. In this paper, we introduce VERTIGO, the first framework for visual preference optimization of camera trajectory generators. Our framework leverages a real-time graphics engine (Unity) to render 2D visual previews from generated camera motion. A cinematically fine-tuned vision-language model then scores these previews using our proposed cyclic semantic similarity mechanism, which aligns renders with text prompts. This process provides the visual preference signals for Direct Preference Optimization (DPO) post-training. Both quantitative evaluations and user studies on Unity renders and diffusion-based Camera-to-Video pipelines show consistent gains in condition adherence, framing quality, and perceptual realism. Notably, VERTIGO reduces the character off-screen rate from 38% to nearly 0% while preserving the geometric fidelity of camera motion. User study participants further prefer VERTIGO over baselines across composition, consistency, prompt adherence, and aesthetic quality, confirming the perceptual benefits of our visual preference post-training. | 28 pages, 10 figures |
| Sliding Mode Control for Safe Trajectory Tracking with Moving Obstacles Avoidance: Experimental Validation on Planar Robots | 2026-04-27 | ShowThis paper presents a unified control framework for robust trajectory tracking and moving obstacle avoidance applicable to a broad class of mobile robots. By formulating a generalized kinematic transformation, we convert diverse vehicle dynamics into a strict feedback form, facilitating the design of a Sliding Mode Control (SMC) strategy for precise and robust reference tracking. To ensure operational safety in dynamic environments, the tracking controller is integrated with a Collision Cone Control Barrier Function (C3BF) based safety filter. The proposed architecture guarantees asymptotic tracking in the presence of external disturbances while strictly enforcing collision avoidance constraints. The novelty of this work lies in designing a sliding mode controller for ground robots like the Ackermann drive, which has not been done before. The efficacy and versatility of the approach are validated through numerical simulations and extensive real-world experiments on three distinct platforms: an Ackermann-steered vehicle, a differential drive robot, and a quadrotor drone. Video of the experiments are available at https://youtu.be/dWcxwum96vk | |
| SceneSelect: Selective Learning for Trajectory Scene Classification and Expert Scheduling | 2026-04-27 | ShowAccurate trajectory prediction is fundamentally challenging due to high scene heterogeneity - the severe variance in motion velocity, spatial density, and interaction patterns across different real-world environments. However, most existing approaches typically train a single unified model, expecting a fixed-capacity architecture to generalize universally across all possible scenarios. This conventional model-centric paradigm is fundamentally flawed when confronting such extreme heterogeneity, inevitably leading to a severe generalization gap, degraded accuracy, and massive computational waste. To overcome this bottleneck, rather than refining restricted model-centric architectures, we propose selective learning, a novel scene-centric paradigm. It explicitly analyzes the characteristics of the underlying scene to dynamically route inputs to the most appropriate expert models. As a concrete implementation of this paradigm, we introduce SceneSelect. Specifically, SceneSelect utilizes unsupervised clustering on interpretable geometric and kinematic features to discover a latent scene taxonomy. A highly decoupled classification module is then trained to assign real-time inputs to these scene categories, and a highly extensible, plug-and-play scheduling policy automatically dispatches the trajectory sequence to the optimal expert predictor. Crucially, this decoupled design ensures excellent generalization capabilities, allowing seamless integration with different off-the-shelf models and robust adaptation across new datasets without requiring computationally expensive joint retraining. Extensive experiments on three public benchmarks (ETH-UCY, SDD, and NBA) demonstrate that our method consistently outperforms strong single-model and ensemble baselines, achieving an average improvement of 10.5%, showcasing the effectiveness of scene-aware selective learning. | This ...This paper has been accepted by ICIC 2026 |
| The Swarm Intelligence Freeway-Urban Trajectories (SWIFTraj) Dataset -- Part II: A Graph-Based Approach for Trajectory Connection | 2026-04-27 | ShowIn Part I of this companion paper series, we introduced SWIFTraj, a new open-source vehicle trajectory dataset collected using a unmanned aerial vehicle (UAV) swarm. The dataset has two distinctive features. First, by connecting trajectories across consecutive UAV videos, it provides long-distance continuous trajectories, with the longest exceeding 4.5 km. Second, it covers an integrated traffic network consisting of both freeways and their connected urban roads. Obtaining such long-distance continuous trajectories from a UAV swarm is challenging, due to the need for accurate time alignment across multiple videos and the irregular spatial distribution of UAVs. To address these challenges, this paper proposes a novel graph-based approach for connecting vehicle trajectories captured by a UAV swarm. An undirected graph is constructed to represent flexible UAV layouts, and an automatic time alignment method based on trajectory matching cost minimization is developed to estimate optimal time offsets across videos. To associate trajectories of the same vehicle observed in different videos, a vehicle matching table is established using the Hungarian algorithm. The proposed approach is evaluated using both simulated and real-world data. Results from real-world experiments show that the time alignment error is within three video frames, corresponding to approximately 0.1 s, and that the vehicle matching achieves an F1-score of about 0.99. These results demonstrate the effectiveness of the proposed method in addressing key challenges in UAV-based trajectory connection and highlight its potential for large-scale vehicle trajectory collection. | |
| Trace2Skill: Distill Trajectory-Local Lessons into Transferable Agent Skills | 2026-04-27 | ShowEquipping Large Language Model (LLM) agents with domain-specific skills is critical for tackling complex tasks. Yet, manual authoring creates a severe scalability bottleneck. Conversely, automated skill generation often yields fragile or fragmented results because it either relies on shallow parametric knowledge or sequentially overfits to non-generalizable trajectory-local lessons. To overcome this, we introduce Trace2Skill, a framework that mirrors how human experts author skills: by holistically analyzing broad execution experience before distilling it into a single, comprehensive guide. Instead of reacting sequentially to individual trajectories, Trace2Skill dispatches a parallel fleet of sub-agents to analyze a diverse pool of executions. It extracts trajectory-specific lessons and hierarchically consolidates them into a unified, conflict-free skill directory via inductive reasoning. Trace2Skill supports both deepening existing human-written skills and creating new ones from scratch. Experiments in challenging domains, such as spreadsheet, VisionQA and math reasoning, show that Trace2Skill significantly improves upon strong baselines, including Anthropic's official xlsx skills. Crucially, this trajectory-grounded evolution does not merely memorize task instances or model-specific quirks: evolved skills transfer across LLM scales and generalize to OOD settings. For example, skills evolved by Qwen3.5-35B on its own trajectories improved a Qwen3.5-122B agent by up to 57.65 absolute percentage points on WikiTableQuestions. Ultimately, our results demonstrate that complex agent experience can be packaged into highly transferable, declarative skills -- requiring no parameter updates, no external retrieval modules, and utilizing open-source models as small as 35B parameters. | Work in Progress |
| Reward-Aware Trajectory Shaping for Few-step Visual Generation | 2026-04-27 | ShowAchieving high-fidelity generation in extremely few sampling steps has long been a central goal of generative modeling. Existing approaches largely rely on distillation-based frameworks to compress the original multi-step denoising process into a few-step generator. However, such methods inherently constrain the student to imitate a stronger multi-step teacher, imposing the teacher as an upper bound on student performance. We argue that introducing \textbf{preference alignment awareness} enables the student to optimize toward reward-preferred generation quality, potentially surpassing the teacher instead of being restricted to rigid teacher imitation. To this end, we propose \textbf{Reward-Aware Trajectory Shaping (RATS)}, a lightweight framework for preference-aligned few-step generation. Specifically, teacher and student latent trajectories are aligned at key denoising stages through horizon matching, while a \textbf{reward-aware gate} is introduced to adaptively regulate teacher guidance based on their relative reward performance. Trajectory shaping is strengthened when the teacher achieves higher rewards, and relaxed when the student matches or surpasses the teacher, thereby enabling continued reward-driven improvement. By seamlessly integrating trajectory distillation, reward-aware gating, and preference alignment, RATS effectively tr |
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.