Time-Series Foundation Models

Follow this topic Ask about this topic

Time-series foundation models learn reusable patterns from large collections of sequences, aiming to support tasks such as forecasting, classification, imputation, and anomaly detection across domains. Common approaches include Transformer models with patch or value tokenization, masked and contrastive pretraining, next-token prediction, and newer recurrent or state-space architectures; training data may mix real series with synthetic ones. Models are used zero-shot or adapted through fine-tuning, in-context examples, covariates, retrieval, and lightweight adapters, with benchmarks such as GIFT-Eval and UCR/UEA testing different capabilities. Transfer can reduce the need for task-specific training, but results vary with data and operating conditions: specialists may still win, and robustness to distribution shifts, long contexts, uncertainty, and fair evaluation remain open challenges.

Latest papers

301 papers69 papers in the last four weeks, up 86% on the four weeks before. 0.4% of all new papers.

Modular Bayesian Foundation Models: Bayesian time series encoders for OOD detection, routing and forecasting
Oct 28, 2026Mame Diarra Toure, Quang Hung Pham, Mohamed Ryad Zemouri +3Time-Series Foundation Models Modular Neural Networks
Department of Mathematics and Statistics McGill University · Hydro-Quebec
Are Seismic Foundation Models More Than Glorified Feature Extractors?
Oct 27, 2026Laura Laurenti, David Liu, Brian Kulis +2Time-Series Foundation Models Self-Supervised Time Series Representation Learning
Institute of Geophysics, ETH Zurich · Swiss Seismological Service, ETH Zurich · Dept. of Computer Science, Boston University +2
Timestral
Oct 27, 2026Adrien Petralia, Alexandre Gavaudan, Artjom Joosen +353Time Series Forecasting Time-Series Foundation Models
TRIDENT-Seq: Profile-Matched Prefix-Future Contrast for Label-Efficient Payment Risk Modeling
Oct 27, 2026Weiyu Han, Haoyu Jiang, Jie Sheng +7Financial Fraud Detection Contrastive Learning
Ant International · East China Normal University · Tongji University +1
Learning from Points and Patches: A Pre-trained Hybrid Attention Transformer for Time Series Analysis
Oct 27, 2026Peiwang Tang, Cheng HE, Yupeng Ye +3Self-Supervised Time Series Representation Learning Time Series Transformers
University of Science and Technology of China, Hefei, Anhui, China · State Key Laboratory of Cognitive Intelligence, Hefei, Anhui, China · Shanghai Black Wing Asset Management, Shanghai, China +1
Self-Supervised Pretraining for PHM: What Transfers, What Does Not, and What Federating It Costs
Oct 26, 2026José Sanz-Durán, David Solís-MartínSelf-Supervised Pre-Training Predictive Maintenance
Department of Computer Science and Artificial Intelligence, Universidad de Sevilla, Seville, Spain
TimeWorld: Towards Action-Conditioned Multimodal World Models for Time Series Forecasting
Oct 26, 2026Hang Ni, Weijia Zhang, Hao LiuTime Series Forecasting Action-Conditioned World Models
The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China
Dynamic Multi-Step Decoding for Time-Series Foundation Models
Oct 25, 2026Birmjune Kim, Hyungju Lee, Hyun Oh SongTime Series Forecasting Time-Series Foundation Models
Department of Computer Science and Engineering, Seoul National University · Interdisciplinary Program in Artificial Intelligence, Seoul National University
Unlocking the Specialization Capability of PPG Foundation Models for Wearable Signals through Structured Pruning and Adaptive Fine-Tuning
Oct 24, 2026Huaming Du, Zhixian Xie, Gaoyang Zhen +6Time-Series Foundation Models Structured Pruning
Southwestern University of Finance and Economics · Jizhi Technology Co., Ltd. · Fuzhou University +2
Sustainability of TSFM Deployment and Pretraining: From Carbon Footprint Analysis to Mitigation via Pruning
Oct 23, 2026Cécile Luc, Nathanaël Lemonnier, Ievgen RedkoTime-Series Foundation Models Energy-Efficient ML
Strategy & Innovation, AI Hub, Schneider Electric, Rueil-Malmaison, France · Strategy & Innovation, AI Hub, Schneider Electric, Grenoble, France
Forecasting Associations, Not Interventions: A Preregistered Causal Audit of Covariate-Aware Time Series Foundation Models
Oct 22, 2026Zong-Han BaiTime-Series Foundation Models Causal Effect Estimation
National Tsing Hua University, Hsinchu, Taiwan.
Zero-Shot Forecasting of In-Situ Soil Moisture and Temperature with a Time-Series Foundation Model
Oct 22, 2026Mohammad Talaei, Mohammad Goudarzi, Maille Todd +2Zero-Shot Time Series Forecasting Spatiotemporal Forecasting
Faculty of Information Technology, Monash University, Australia · School of Earth, Atmosphere and Environment, Monash University, Australia · The Environmental Informatics Hub, Data Science and AI, Monash University, Australia
Random Initialization as a Baseline for Time Series Foundation Models
Oct 22, 2026Pinar Sungu Isiacik, Georgiana IfrimTime-Series Foundation Models Self-Supervised Pre-Training
University College Dublin, Dublin, Ireland
Rethinking Agentic Model Orchestration for Time Series Forecasting
Oct 21, 2026Tong Guan, Tianxiang Zhan, Shirui Pan +1Agentic Time Series Forecasting Time Series Forecasting
Griffith University · Zhejiang University
Understanding the Inner Workings of a State-Space Foundation Model for Limit Order Books
Oct 20, 2026Valentin Mohl, Alexandre Bismuth, Aramis Marti-Shahandeh +6Limit Order Book Modeling State Representation Learning
British Open-ended Learning and Discovery Lab (BOLD), Department of Computer Science, University of Oxford, Oxford, United Kingdom · Harvard University, Cambridge, MA, United States · British Open-ended Learning and Discovery Lab (BOLD), Department of Engineering Science, University of Oxford, Oxford, United Kingdom +3
Turning TS2Vec into a Foundation Model for Time Series Classification
Oct 20, 2026Vasilii Feofanov, Zhenghan Tai, Andrei Ermakov +1Time Series Classification Time-Series Foundation Models
NL · University of Toronto · Université de Montréal Mila - Quebec AI Institute
Parameter-Efficient Adaptation of Time Series Foundation Models for Electrical Load Forecasting
Oct 20, 2026Aliullah, Julian Oelhaf, Edouard Fouché +4Electricity Load Forecasting Time-Series Foundation Models
Pattern Recognition Lab, Friedrich-Alexander-Universität Erlangen-Nürnberg · Siemens AG, Smart Infrastructure · Department of Electrical Engineering, Media and Computer Science, Ostbayerische Technische Hochschule Amberg-Weiden
Foundation-Style Pretraining for Scalable and Transferable Distributed PV Forecasting
Oct 20, 2026Dayin Chen, Xiaodan Shi, Xin Xu +6Solar Power Forecasting Time-Series Foundation Models
Department of Building Environment and Energy Engineering, The Hong Kong Polytechnic University, Hong Kong SAR, China · International Centre of Urban Energy Nexus, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China · Ningbo Institute of Digital Twin, Eastern Institute of Technology, Ningbo, 315200, China +2
WhatIf-Bench: Benchmarking Intervention-Aware Time Series Forecasting
Oct 19, 2026Arjun Ashok, Mihir Parmar, Palash Goyal +6Time Series Forecasting Time Series Forecasting with Covariates
Google · Mila - Quebec AI Institute · Université de Montréal +1
Beyond Static Leaderboards: Capability-Focused Evaluation with Structural Diversity for Time Series Foundation Models
Oct 19, 2026Hongyin Zhang, Mingyi Xiang, Rongzhao Chen +7Forecasting Benchmarks Time-Series Foundation Models
School of Software, BNRist, Tsinghua University · Timecho Ltd, Beijing, China · Harbin Institute of Technology
From Retrieval to Recognition:Knowledge-Enhanced Post-training of Time Series Foundation Models for Classification
Oct 19, 2026Zhen Liu, Yucheng Wang, Yingpeng Du +5Time Series Classification Time-Series Foundation Models
Nanyang Technological University, Singapore · Institute of Advanced Intelligence and Computing (IAIC), A*STAR, Singapore · AI Center of Excellence for Manufacturing (AIMfg), A*STAR, Singapore +2
Adversarial Post-Training of Foundation Models for High-Frequency Trading
Oct 19, 2026Alexandre Bismuth, Kang Li, Aramis Marti-Shahandeh +3Limit Order Book Modeling Adversarial Training
British Open-ended Learning and Discovery Lab (BOLD), Department of Computer Science, University of Oxford, Oxford, United Kingdom · Harvard University, Cambridge, MA, United States · British Open-ended Learning and Discovery Lab (BOLD), Department of Statistics, University of Oxford, Oxford, United Kingdom +1
World Models, Trees and Foundation Models for Retail Forecasting: Evidence from Four Public Data Sets
Oct 18, 2026Karthik Iyer, Madan DabbeeruTime Series Forecasting Time Series Forecasting with Covariates
Morpheus-2.0: Multiscale Probabilistic Forecasting with Dual-Axis Attention
Oct 16, 2026Aakriti, Manoj Cherukumalli, Prathamesh Patil +5Time Series Forecasting Probabilistic Time Series Forecasting
Perceive, Solve, Reason: Progressive Capability Formation in Time Series-LLMs
Oct 16, 2026Xin Qiu, Junlong Tong, Wei Zhang +1Time Series Reasoning Time Series Language Modeling
EIT-NLP Lab, Eastern Institute of Technology, Ningbo · Zhejiang University · Shanghai Jiao Tong University
What Do Foundation Models for Market Microstructure Learn? A Controlled Study with Synthetic Limit Order Books
Oct 14, 2026Fernando AceroLimit Order Book Modeling Time-Series Foundation Models
J.P. Morgan QTR AI Research, United Kingdom
TimeRSI: Towards Self-Improved Agentic Deployment for Time-Series Foundation Models
Oct 14, 2026Haoxin Liu, Deqing Fu, Lucheng Fu +4Agentic Time Series Forecasting Time Series Forecasting
Georgia Institute of Technology · University of Southern California · Samay AI
TEX-MEX: Similarity-Guided Temporal Contrastive Learning for Time Series
Oct 14, 2026Aditya Radhakrishnan, Maxwell A. Xu, Wanting Mao +3Self-Supervised Time Series Representation Learning Contrastive Learning
Health Care Engineering Systems Center, University of Illinois Urbana-Champaign · Google · Department of Computer Science, University of Memphis
Test-Time Forecasting: Recurrent Depth Enables In-Context Dynamics Prediction
Oct 14, 2026Jeffrey B. Lai, John Quinn, Anthony Bao +1Dynamical Systems In-Context Learning
Oden Institute University of Texas at Austin · Department of Physics University of Texas at Austin
GIFT-Eval-CTX: A Benchmark For Contextual Forecasting Model Evaluation
Oct 13, 2026Xu Liu, Hanshu Yan, Taha Aksu +6Time Series Forecasting Forecasting Benchmarks
Moirai Team, Salesforce AI Research

Oct 28, 2026J/K move · Enter open · S save

Modular Bayesian Foundation Models: Bayesian time series encoders for OOD detection, routing and forecasting

Mame Diarra Toure, Quang Hung Pham, Mohamed Ryad Zemouri, David Stephens+2

Department of Mathematics and Statistics McGill University · Hydro-Quebec

Time-Series Foundation Models Modular Neural Networks Probabilistic Time Series Forecasting Expert Routing Bayesian Neural Networks OOD Detection

Open paper AskPDF arXiv

Industrial systems continuously generate time series that vary in sampling frequency, scale, dynamics, and operating regime. Consequently a single time-series foundation model may not be able to represent all regimes effectively and retraining a large model for every new dataset or regime shift is computationally expensive. Modular foundation models address this issue by reusing specialised encoders trained on diverse datasets and by allowing the expansion of capabilities through the addition of new experts. This approach requires compact experts to minimize storage cost and a proper selection mechanism of experts for unseen data. We propose a modular foundation model framework that combines end-to-end Bayesian encoder training with uncertainty based routing. Using Singular Bayesian Neural Networks (SBNNs), we build a collection of compact Bayesian encoders whose epistemic uncertainty guides label-free selection. At the selected ranks, our Bayesian collection uses 63.8% less storage than its deterministic counterpart. We evaluate this framework on time-series data through distribution shift, encoder selection, and probabilistic forecasting. Across seven incoming datasets, uncertainty-based routing is competitive with distance-based routing in terms of forecast MSE, while reducing average ECE from 0.086 to 0.072 and average NLL from 0.907 to 0.836. It also increases PICP90 from 82.28% to 89.03%. In a separate model-family comparison, we evaluate the SBNN pipeline against full-rank Bayes by Backprop, deep ensembles, and MC Dropout. The SBNN pipeline achieves an average MSE of 0.390, comparable with 0.396 for deep ensembles, 0.389 for MC dropout, and 0.443 for the full-rank Bayesian model. It uses 63.9% fewer parameters than MC dropout, 81.9% fewer than the full-rank model, and 92.8% fewer than the deep ensemble.