Temporal Distribution Shift

Follow this topic Ask about this topic

Temporal distribution shift concerns changes in the data a model encounters over time, such as new time-series regimes, evolving language, or changing user and device populations. Researchers measure and detect drift, then address it with online or continual learning, adaptive normalization, memory and expert models, retrieval, and calibration. Evaluation increasingly uses chronological, leakage-free tests and tracks accuracy, calibration, forgetting, and deployment costs across time. A central tension is how to adapt quickly to real change without overreacting to noise, forgetting useful patterns, or relying on future information that would not be available in deployment.

Latest papers

43 papers2 papers in the last four weeks, down 67% on the four weeks before. 0.0% of all new papers.

When to Trust Memory: Retrieval-Guided Probabilistic Spatiotemporal Forecasting under Distribution Shift
Oct 23, 2026Haochen Lv, Jianhao Zhang, Zhichen Lei +5Spatiotemporal Forecasting Probabilistic Forecasting
School of Computer Science and Technology, Beijing Jiaotong University, Beijing, China. · Key Laboratory of Big Data and Artificial Intelligence in Transportation, Ministry of Education, Beijing, China. · Beijing Key Laboratory of Traffic Data Mining and Embodied Intelligence, Beijing, China.
LLM Evaluation is a Forecasting Problem
Oct 12, 2026Rahul Thomas, Adam Block, Micah GoldblumLanguage Model Performance Prediction LLM Evaluation
Department of Electrical Engineering, Columbia University. · Department of Computer Science, Columbia University.
RAEGL: Risk-Aware Evidence-Gated Learning for Selective Contextual Routing under Temporal Shift
Sep 27, 2026Yifan GuoDistribution Shift Selective Prediction
Stony Brook University
Model-Agnostic Online Certificate-Driven Calibration for Time Series Forecasting Under Distribution Shift
Sep 25, 2026Chenfeng Huang, Zixuan Ma, George MichailidisOnline Calibration Time Series Forecasting
Department of Statistics and Data Science, University of California, Los Angeles, Los Angeles, California, USA
AlphaEarth distinguishes cities but compresses urban variation
Sep 24, 2026Andrew RenningerGeospatial Foundation Models Geospatial Representation Learning
School of Geographical & Earth Sciences, University of Glasgow
Temporal Generalization and Explanation Stability of Control Flow Graph Neural Networks for Malware Detection
Sep 21, 2026Md. Asif Sajeed, Md. Nazrul Islam Mondal, Md Ashraful Hossen AkashGraph Neural Networks Distribution Shift
Department of Computer Science and Engineering, Rajshahi University of Engineering & Technology, Rajshahi, 6204, Bangladesh
The Anatomy and Boundary of Adaptation under Temporal Tabular Shift
Sep 10, 2026Tianyu Wang, Xi Vincent Wang, Lihui Wang +2Tabular Foundation Models Distribution Shift
Department of Production Engineering, KTH Royal Institute of Technology, Stockholm, Sweden · Global Institute of Future Technology, Shanghai Jiao Tong University, Shanghai, China
When Does Online Adaptation Pay on the Edge? A Leakage-Free Evaluation of Warmup, Learning-Rate Selection, and Resource Trade-offs for Time-Series Forecasting
Sep 1, 2026Takumi Fujimoto, Hiroaki NishiMultivariate Time Series Forecasting Learning Rate Scheduling
School of Science for Open and Environmental Systems, Graduate School of Science and Technology, Keio University, Yokohama, Japan
When Writing Style Drifts: Benchmarking Authorship Verification under Distribution Shifts in Genre, Time and the AI-Era
Aug 18, 2026Lotta Kiefer, Brisca Balthes, Christoph Leiter +3Distribution Shift Robustness Authorship Attribution
University of Technology Nuremberg (UTN)
When Can You Correct Distribution Drift in Temporal Graph Generation? A Sharpening--Drift Tension and an Impossibility for Observation-Based Correction
Jul 27, 2026Tianpeng Li, Xuan Guo, Wenjun Wang +2Flow Matching Distribution Shift
College of New Media and Communication, Tianjin University · School of Artificial Intelligence, Tianjin University · School of Cyberspace Security,Hangzhou Dianzi University
Parameter-Efficient Vision-Language Adaptation with Continuous Metadata Conditioning for Animal Re-Identification
Jul 10, 2026Anil Osman Tur, Tonje Knutsen Sordalen, Kim Tallaksen Halvorsen +1Animal Re-Identification Distribution Shift
Department of Computer Science, University of Verona, Verona, 37134, Italy · Institute of Marine Research, Nye Flødevigveien 20, 4817 His, Norway · University of Agder, Centre for Coastal Research, Grimstad, 4604, Norway
Temporal Knowledge Graph Forecasting under Distribution Shifts: A Synthetic Evaluation
Jul 10, 2026Konrad Özdemir, Julia Gastinger, Lukas Kirchdorfer +1Distribution Shift Temporal Link Prediction
Data and Web Science Group, University of Mannheim, Germany · SAP Signavio, Walldorf, Germany
Drift Happens: An Empirical Study of Neural Architecture Robustness to Temporal Distribution Shift
Jul 7, 2026Robin Holzinger, Riccardo CollettiDistribution Shift Robustness Neural Network Generalization
Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, USA
Domain-Adaptive Climate Downscaling Under Temporal Distribution Shift
Jul 6, 2026Shuochen Wang, Nishant Yadav, Auroop R. GangulyDistribution Shift Robustness Domain Adaptation
Sustainability and Data Sciences Laboratory, Northeastern University · Microsoft · AI for Climate and Sustainability, The Institute for Experiential AI, Northeastern University
When Multi-Sensor Fusion Fails to Generalize: Cattle Posture Classification Under Animal-Level and Temporal Distribution Shift
Jun 23, 2026Leutrim Uka, Severino Pinto, Gundula Hoffmann +1Distribution Shift Robustness Distribution Shift
Institute of Computer Science, University of Potsdam, Potsdam, Germany. · Department of Sensors and Modelling, Leibniz Institute for Agricultural Engineering and Bioeconomy - ATB, Potsdam, Germany. · Department of Data Science in Bioeconomy, Leibniz Institute for Agricultural Engineering and Bioeconomy - ATB, Potsdam, Germany.
Betting on Moments: Legendre Jumper Martingales for Online Exchangeability Testing
Jun 18, 2026Johan Hallberg SzabadváryConformal Prediction Distribution Shift Detection
Department of Mathematics, Stockholm University
EPM-JEPA: Operator-Side Experience Modulation in JEPA-Family World Models
Jun 11, 2026Vedant PandyaTest-Time Adaptation World Model Learning
School of Artificial Intelligence and Data Engineering (SAIDE) Indian Institute of Technology Jodhpur
TEDD: Robust Detection of Unstable Temporal Features
Jun 10, 2026Ricardo Ribeiro Pereira, Bruno Casal Laraña, Nádia Soares +1Distribution Shift Robustness Distribution Shift Detection
DCC/FCUP University of Porto Porto, Portugal · Feedzai Porto, Portugal · Mindera Porto, Portugal
Machine-Learning Emulation of Satellite Greenhouse Gas Retrievals: Stability over Time
Jun 8, 2026Nugzar Gognadze, Motonobu Kanagawa, Yu Someya +1Surrogate Modeling Distribution Shift
Data Science Department, EURECOM, 450 Route des Chappes, Biot, 06410, France
Empirical Transfer Operators and Finite-Sample Change Detection for Noisy Expanding Interval Maps
Jun 5, 2026Aparna RajputDynamical Systems Distribution Shift Detection
FlashbackCL: Mitigating Temporal Forgetting in Federated Learning
Jun 2, 2026Mubarak A. Ojewale, Adriana E. Chis, Jorge M. Cortes-Mendoza +2Distribution Shift Federated Continual Learning
Cloud Competency Centre, National College of Ireland, Dublin, Ireland
Regime-Arrival Uncertainty in Generalization Bounds under Distribution Shift
Jun 1, 2026Prince PoudelDistribution Shift Robustness Statistical Learning Theory
Independent researcher.
DistMatch: Adaptive Binning via Distribution Matching for Robust Sequential Conformal Prediction
May 30, 2026Enver Menadjiev, Jihyeon Seong, Jisu Yeo +1Distribution Shift Robustness Conformal Prediction
Korea Advanced Institute of Science and Technology, South Korea · INEEJI, South Korea
dashi: A Python library for Dataset Shift Characterization to Support Trustworthy AI Development and Deployment
May 29, 2026David Fernández-Narro, Pablo Ferri, Ángel Sánchez-García +2Distribution Shift Detection Temporal Distribution Shift
Biomedical Data Science Lab, Instituto Universitario de Tecnologías de la Información y Comunicaciones, Universitat Politècnica de Valéncia, Camino de Vera s/n, Valencia 46022, España
ROAR: Retrieval Opportunity-Aware Refinement for Zero-Shot Time Series Forecasting
May 24, 2026Jinjin Chi, Lei Feng, Yiming Wang +4Zero-Shot Time Series Forecasting Time Series Forecasting
College of Computer Science and Technology, Jilin University, Changchun, China · College of Computing and Data Science, Nanyang Technological University, Singapore
Temporal Concept Drift in Legal Judgment Prediction: Neural Baselines Across Three Epochs of Ukrainian Court Decisions
May 23, 2026Volodymyr OvcharovLanguage Model Pretraining Legal NLP
Dynamic TMoE: A Drift-Aware Dynamic Mixture of Experts Framework for Non-Stationary Time Series Forecasting
May 20, 2026Jiawen Zhu, Shuhan Liu, Di Weng +1Concept Drift Mixture of Experts
School of Software Technology, Zhejiang University, Ningbo, China · State Key Lab of CAD&CG, Zhejiang University, Hangzhou, China
DeRegiME: Deep Regime Mixtures for Probabilistic Forecasting under Distribution Shift
May 19, 2026Kieran Wood, Stefan Zohren, Stephen J. RobertsDistribution Shift Robustness Time Series Forecasting
Machine Learning Research Group, University of Oxford · Oxford-Man Institute, University of Oxford
L-Drive: Beyond a Single Mapping-Latent Context Drives Time Series Forecasting
May 18, 2026Fan Zhang, Shijun Chen, Hua WangMultivariate Time Series Forecasting Latent Dynamics Modeling
Shandong Technology and Business University, Yantai, Shandong, China · Ludong University, Yantai, Shandong, China
Attention Dispersion in Dynamic Graph Transformers: Diagnosis and a Transferable Fix
May 15, 2026Jinhao Zhang, Kangfei Zhao, Qiuhao Zeng +1Distribution Shift Robustness Temporal GNNs
Beijing Institute of Technology · University of Toronto · Hong Kong Baptist University

Oct 23, 2026J/K move · Enter open · S save

When to Trust Memory: Retrieval-Guided Probabilistic Spatiotemporal Forecasting under Distribution Shift

Haochen Lv, Jianhao Zhang, Zhichen Lei, Yang Jiang+4

School of Computer Science and Technology, Beijing Jiaotong University, Beijing, China. · Key Laboratory of Big Data and Artificial Intelligence in Transportation, Ministry of Education, Beijing, China. · Beijing Key Laboratory of Traffic Data Mining and Embodied Intelligence, Beijing, China.

Spatiotemporal Forecasting Probabilistic Forecasting Distribution Shift Robustness Temporal Distribution Shift Retrieval-Augmented Time Series Forecasting

Open paper AskPDF arXiv

Spatiotemporal forecasting plays an important role in real-world applications, where probabilistic approaches provide uncertainty estimates for decision-making. Real-world systems are inherently non-stationary, leading to distribution shifts between training and test data. However, existing probabilistic methods mainly focus on modeling data distributions under standard settings, while distribution shift methods often handle such shifts in a coarse manner, failing to characterize shift degree and adapt prediction strategy accordingly. As a result, they often produce miscalibrated and unreliable forecasts. To address this issue, we propose REMIND, a retrieval-guided probabilistic framework for spatiotemporal forecasting under distribution shift. REMIND leverages retrieval to estimate how well current observation are supported by historical patterns, enabling adaptive prediction under different shift regimes. It employs spatiotemporal retrieval to construct a memory bank and introduces a dual branch architecture, combining a memory-calibrated probabilistic branch for mild shifts with a deviation-aware probabilistic branch for severe shifts. A shift-adaptive gate, guided by retrieval similarity and memory dispersion, balances the two branches according to shift degree. We further design a hierarchical Gaussian mixture to capture heterogeneous uncertainty over the dual branch predictions, and introduce a memory perturbation training scheme to improve generalization. Experiments on real-world datasets with distribution shifts demonstrate that REMIND consistently outperforms state-of-the-art methods in both forecasting accuracy and probabilistic prediction quality.