WaveHiTS: Wavelet-Enhanced Hierarchical Model for Wind Direction Nowcasting in Inner Mongolia

A revised arXiv paper introduces WaveHiTS, a wavelet-enhanced hierarchical time series model designed for short-term wind direction forecasting in eastern Inner Mongolia. The approach targets common difficulties in directional data, including circular values, multi-step error accumulation, and complex meteorological interactions. It was cross-listed on arXiv's cs.LG and cs.AI categories.

Optimal Transport Framework Proposed for Unsupervised Industrial Anomaly Detection

A new arXiv paper presents an anomaly detection framework built on optimal transport, aimed at spotting deviations from normal behavior in industrial time-series data. The authors position the approach as unsupervised and computationally efficient, targeting Industry 4.0 monitoring use cases. Details on datasets, baselines, and evaluation results are not available from the abstract alone.

CodeTS Generates Time Series from Text via Executable Code

A new arXiv paper introduces CodeTS, a method that turns natural-language descriptions into time series by generating and running executable code rather than sampling outputs directly. This design makes the resulting synthetic data verifiable and suited to scenarios where real observations are scarce or expensive to collect. The work appears in the cs.LG and cs.AI listings.

LLaTSA: general-purpose transient stability analysis aligned with LLMs

A new arXiv paper introduces LLaTSA, a method that adapts large language models to predict dynamic trajectories for power-system transient stability analysis. Most existing data-driven predictors are tied to a specific grid and need retraining when network topology or generation mix changes, whereas this approach aims to generalize across configurations. The work is a cross-listing on cs.AI and falls under research rather than a released product.

Study examines parameter-efficient tuning of language models for time-series forecasting

A new arXiv paper investigates how pretrained language models can be adapted for univariate time-series forecasting using parameter-efficient transfer learning. The authors focus on identifying which design decisions matter most for effective transfer between text and numerical sequences. The work is a cross-listed submission to arXiv's machine learning category.

Explainable GNN Framework Targets Component-Level Anomaly Diagnosis in Industrial Systems

A revised arXiv paper proposes a graph neural network framework for diagnosing which specific sensor or component is responsible for anomalies in multivariate industrial time series. The work aims to go beyond simple anomaly detection by making the source of the fault interpretable. It targets reliability and safety monitoring in complex, multi-sensor process environments.

arXiv paper proposes Hilbert-valued framework for explaining time-dependent model outputs

A new preprint introduces a decomposition method that extends feature-attribution explanations from single-number predictions to functional or multivariate outputs, such as demand forecasts that vary over time. The approach works in a Hilbert space so that the influence of each input feature can be separated across the whole output trajectory rather than summarized by one score. The authors position it as a general framework for settings where model predictions are curves or vectors instead of scalars.

Paper Proposes Attention-Based Method for Multivariate Time Series Anomaly Detection

A revised arXiv paper introduces a technique that flags anomalies in multivariate time series by tracking shifts in cross-channel dependencies rather than only large amplitude changes. The authors illustrate the idea with autonomous driving, where a steering command can look internally consistent yet no longer match the resulting vehicle behavior. The work appears on arXiv under cs.AI and cs.LG as a cross-listing update.

FlowTSFM: Turning Encoder Depth into Quantile Transport

A new arXiv preprint introduces FlowTSFM, an approach for encoder-based time series foundation models that assigns a predictive role to intermediate Transformer layers instead of supervising only the final forecast. The method recasts encoder depth as a form of quantile transport, according to the abstract. The announcement provides only the opening portion of the paper's abstract, so full details of the architecture and evaluation are not yet available in this report.

TypiCore: Hybrid Active Query Strategy for Class-Incremental Time Series Learning

A new arXiv paper introduces TypiCore, a method that combines active learning queries with class-incremental learning for time series data. The approach targets settings such as healthcare and manufacturing, where models must adapt to distribution shifts as new classes arrive over time. The work is posted as a cross-listing replacement on arXiv and has not yet undergone peer review.

Survey Maps Probabilistic Forecasting Methods for Time Series and Spatiotemporal Data

A new arXiv survey examines how probabilistic forecasting methods have developed across time series and spatiotemporal research. The authors argue the field has grown fragmented, with statistical modeling, machine learning, and deep generative approaches advancing largely in parallel. The paper aims to organize this landscape and connect the differing methodological traditions used for forecasting under uncertainty.

Tabby: Open Pretraining Recipe Released for Time Series Foundation Models

Researchers introduce Tabby, a long-context probabilistic foundation model designed for time series data, built on an encoder-only patch Transformer architecture. The release includes a fully open account of the pretraining process, covering the decisions and components behind the model's construction. The work aims to make time series foundation model development more reproducible and accessible.

Study Asks Whether Gradient Boosting Models Fit Intermittent Demand Forecasting

A new arXiv paper examines how well gradient boosting methods handle demand forecasting for products with intermittent, sporadic sales patterns. Such cases are difficult because standard forecasting approaches tend to struggle when demand is irregular. The work assesses whether gradient boosting is an appropriate tool in this setting.

WaVeFuse Model Combines Wavelet Denoising and Attention for Equity Index Forecasting

A new arXiv paper introduces WaVeFuse, a hybrid deep learning approach for forecasting stock market indices. The method targets three issues in existing models: noise from OHLCV data leaking into derived technical indicators, treating all channels the same during multi-scale decomposition, and mismatched frequency signals. It applies channel-wise wavelet denoising with vertical attention fusion to adapt across market regimes.

Survey Reviews Frequency-Domain Methods for Time Series Analysis

A new survey examines how frequency-domain techniques are being applied to time series analysis, contrasting them with conventional time-domain approaches. It outlines unresolved theoretical and practical problems in the field and discusses benchmarks and open research directions.

arXiv Paper Proposes Horizon-Specific Expert Fusion for Solar Power Forecasting

A new arXiv preprint describes a hierarchical ensemble method for short-term photovoltaic power forecasting. The approach pairs temporal neural models with expert fusion that adapts to the forecast horizon, reflecting that regular solar cycles and weather-driven variation matter differently at different lead times. The work targets improved accuracy in solar generation forecasts.

Paper applies minimum description length to temporal misalignment in multichannel time-series classification

A new arXiv preprint examines how unsynchronized sensor streams — caused by latency, clock drift, or preprocessing — degrade multichannel time-series classification. The authors propose using the minimum description length principle to diagnose and characterize these relative delays. The work targets a setting where most existing methods assume channels are already aligned.

TimeThink: Method Aims to Improve Compositional Reasoning in Time-Series LLMs

A new arXiv paper introduces TimeThink, a technique intended to help time-series multimodal large language models reason more compositionally. The authors note that such models often struggle to capture dynamic temporal patterns when answering questions. The work focuses on eliciting stronger reasoning behavior from these models rather than treating forecasting as pure pattern matching.

Paper examines distilling synthetic data for time series foundation models

A new arXiv preprint looks at how time series foundation models are pretrained on artificially generated trajectories where the underlying data-generating process is known. The work focuses on distillation methods rather than the conventional loss-based pretraining objectives that compare model outputs against targets. It aims to improve how these models learn from synthetic time series data.

T-SMART: Mechanism-Level Attribution for Tool-Augmented Time-Series QA

A new arXiv paper introduces T-SMART, a method that attributes how tool-augmented language models arrive at answers for time-series question answering. The work targets a known weakness: LLMs handle poorly the case where numerical signals are turned into text and require explicit computation. It offers analysis at the level of internal mechanisms rather than just final outputs, though the abstract provided is truncated.

Unified benchmark targets multimodal time series forecasting with heterogeneous context

Researchers present a new benchmark for time series forecasting that moves beyond purely numeric data to include contextual signals such as text and other modalities. They argue existing multimodal benchmarks are limited in both the volume of data and the range of context they cover. The work is posted as an arXiv preprint in cs.AI and cs.LG.

Prototype-Regularized Graph Structure Learning for Multivariate Time Series Anomaly Detection

A new arXiv paper introduces GSLAD, an unsupervised method for detecting anomalies in multivariate time series. The approach focuses on changes in the structural relationships between variables, which often appear before individual readings deviate, and regularizes learned graph structures with prototypes. The authors argue this addresses a gap in forecasting- and reconstruction-based detection methods.

Shapelet-Based Distance Measure Aims to Improve Multi-Source Transfer Learning for Time Series

A new arXiv preprint proposes selecting source datasets for time series classification by measuring similarity through shapelets, the discriminative subsequences that characterize time series patterns. The authors argue that transfer learning helps overcome limited labeled data in deep learning, but its usefulness hinges on picking appropriate source datasets. Their method is presented as an alternative to conventional transferability estimation, which the paper describes as computationally expensive.

BiHDTrans: Binary Hyperdimensional Transformer for Edge Time Series Classification

Researchers propose BiHDTrans, a transformer variant that uses binary hyperdimensional computing to classify multivariate time series from IoT sensors. The design targets resource-constrained edge devices, where large data volumes and limited compute make standard models impractical. It is presented as an arXiv preprint focused on balancing efficiency with classification accuracy.

Google Research releases TimesFM-3, a 330M-parameter forecasting model

Google Research has introduced TimesFM-3, a model designed to forecast time series by combining historical data with related signals and known upcoming events such as promotions or weather. It generates all future time points at once rather than predicting one step at a time, and has 330 million parameters.

SurF: A Generative Model for Multivariate Irregular Time Series Forecasting

Researchers introduce SurF, a generative model designed for multivariate event streams that are sampled at irregular intervals. The work argues that tokenization-based approaches struggle when the gaps between events span orders of magnitude, and proposes an alternative suited to such data. The paper is a revised arXiv submission in machine learning.

PeriodicCALM: real-time anomaly detection for cyclostationary data streams

A new arXiv preprint presents PeriodicCALM, a framework for detecting anomalies on the fly in cyclostationary data streams, whose statistical properties vary periodically over time. Rather than relying on classical stationary assumptions, the algorithm adapts to these recurring temporal patterns as it monitors live data for deviations. The work was posted to arXiv's machine learning listing as a cross-listed submission.

Study examines trade-off between forecast horizon length and learnability in autoregressive models

A new study investigates how far into the future autoregressive models should be trained when forecasting dynamical systems. The authors identify a trade-off between predictive performance and learnability as the training horizon grows, suggesting an optimal horizon exists. The findings offer practical guidance for selecting prediction horizons during model training.

KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning

Researchers introduce KairosAgent, an agentic approach to forecasting time series that span multiple domains and data types. The framework is designed to pair careful numerical analysis with semantic interpretation and to fuse information from different modalities. The paper is a revised version (v2) recently updated on arXiv.

Online Learning of Scale Parameters in Score-Driven Filters

A new research paper addresses how to learn the gain, the scale parameter that multiplies the scaled log-likelihood score in score-driven filters, directly online. Rather than fixing this coefficient beforehand, the method treats each admissible gain as selecting a reachable next state given the current state and the realized scaled score. This allows the filter's update step to adapt during operation.

Unsupervised Anomaly Detection Framework for Spacecraft Telemetry Uses Adaptive EVT Thresholding

Researchers have introduced an unsupervised framework for detecting anomalies in spacecraft telemetry that does not rely on labeled historical anomalies or lengthy warm-up periods, addressing common barriers to real-world deployment. The approach uses structure-aware modeling combined with adaptive Extreme Value Theory (EVT) thresholding to determine when telemetry readings should be flagged. The authors position the method as ready for operational use in settings where annotated failure data is scarce.

Time-Series Foundation Model Benchmarks Still Reflect Pretraining Familiarity on Later Hold-Outs

A new study questions whether time-series foundation models can be fairly evaluated using test data collected after their pretraining cutoff. It finds that even a temporally later, contamination-free hold-out does not fully isolate genuine generalization, as familiarity with the underlying data distribution absorbed during pretraining persists. The result suggests the field needs evaluation practices that go beyond simply withholding recent data.

Jump-Diffusion Framework Introduced for Generating Irregularly Sampled Time Series

A research paper presents a method for training generative models on continuous-time data that is recorded unevenly and out of sync across sources. The approach builds on generator matching and can represent trajectories with sudden jumps rather than only smooth paths, backed by closed-form expressions for diffusion components. It could be useful in domains where measurements arrive at irregular intervals, such as healthcare monitoring or sensor networks.

Benchmarking Information-Theoretic Metrics for Multimodal Time-Series Forecasting

A new arXiv paper examines how to measure whether text annotations actually improve a forecasting model's predictions when paired with time-series data. The authors propose benchmarking information-theoretic metrics intended to quantify how much a given text input contributes to forecast accuracy. The study targets multimodal forecasting pipelines that blend numerical series with textual context.

LoaDiff: Conditional Generation of Electricity Consumption Time Series

A new arXiv preprint introduces LoaDiff, a method for conditionally generating residential electricity consumption time series. The work is motivated by the energy transition, where distributed generation, electrified appliances and demand-response programs are shifting how households use power. The authors argue that granular synthetic consumption data can support energy analytics; the abstract is truncated in this report.

CryptoL Framework Targets Scale Imbalance in Cryptocurrency Forecasting

A new arXiv paper introduces CryptoL, a unified framework for forecasting multivariate cryptocurrency time series. The work addresses cross-asset scale differences, non-stationary market dynamics, and dependencies among open, high, low, and close price variables. It proposes physics-informed constraints to mitigate these issues in financial prediction.

IBM releases Granite Time Series PatchTST-FM-r2 model with commercial-friendly license

IBM has published a revised edition of its Granite time series foundation model, PatchTST-FM-r2, on Hugging Face. Built on the Patch Time Series Transformer architecture, the model targets forecasting workloads and is offered under licensing terms that permit commercial use.

WHY IT MATTERS ↘A commercially licensed, openly available forecasting foundation model lowers the cost and legal friction of adopting time-series AI in production, an area where enterprises have mostly faced proprietary or research-restricted options. The rapid revision also signals that vendors are competing on maintained, enterprise-ready time-series models rather than one-off releases.