ICML 2026 accepted papers on Time series
138 of the 6,341 papers accepted at ICML 2026 (2.2%) are on Time series, compared with 2.1% at ICML 2025. They fall into 4 topics.
See ICML 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Foundation models for time series · 59 papers
- TwinWeaver: An LLM-Based Foundation Model Framework for Pan-Cancer Digital Twins (2 citations)
- Time Series, Vision, and Language: Exploring the Limits of Alignment in Contrastive Representation Spaces (1 citation)
- Adaptive Time Series Reasoning via Segment Selection
- Beyond Extrapolation: Knowledge Utilization Paradigm with Bidirectional Inspiration for Time Series Forecasting
- BioFormer: Rethinking Cross-Subject Generalization via Spectral Structural Alignment in Biomedical Time-Series
- Byte Pair Encoding for Efficient Time Series Forecasting
- CoGenCast: A Coupled Autoregressive–Flow Generative Framework for Time Series Forecasting
- Divide and Contrast: Learning Robust Temporal Features without Augmentation
- From Text to Forecasts: Bridging Modality Gap with Temporal Evolution Semantic Space
- Giving Sensors a Voice: Multimodal JEPA for Semantic Time-Series Embeddings
- HEARTS: Benchmarking LLM Reasoning on Health Time Series
- Invariant Representation Learning for Source-Free Time Series Forecasting with LLM-Centric Proxy Denoising
and 47 more in the interactive map.
Causal and generative time series · 37 papers
- Beyond Model Ranking: Predictability-Aligned Evaluation for Time Series Forecasting
- Causal Discovery for Irregularly Time Series with Consistency Guarantees
- CombinationTS: A Modular Framework for Understanding Time-Series Forecasting Models
- Continuous-Time Piecewise-Linear Recurrent Neural Networks
- ConTSG-Bench: A Unified Benchmark for Conditional Time Series Generation
- DAG: A Dual Correlation Network for Time Series Forecasting with Exogenous Variables
- Diff-MN: Diffusion Parameterized MoE-NCDE for Continuous Time Series Generation with Irregular Observations
- Dynamic TMoE: A Drift-Aware Dynamic Mixture of Experts Framework for Non-Stationary Time Series Forecasting
- Ellipsoidal Time Series Forecasting
- From Observations to States: Latent Time Series Forecasting
- Hierarchical ODE: Learning Continuous-Time Physical Prototypes for Early Link Failure Detection
- How (Not) to Hybridize Neural and Mechanistic Models for Epidemiological Forecasting
and 25 more in the interactive map.
Long-term forecasting · 35 papers
- Dynamic Relational Priming Improves Transformer in Multivariate Time Series (1 citation)
- Benchmarking Physics-Informed Time-Series Models for Operational Global Station Weather Forecasting
- Bridging Time and Frequency: A Joint Modeling Framework for Irregular Multivariate Time Series Forecasting
- Delving into Non-Exchangeability for Conformal Prediction in Graph-Structured Multivariate Time Series
- DropoutTS: Sample-Adaptive Dropout for Robust Time Series Forecasting
- Generalizing Multi-Scale Time-Series Modeling with a Single Operator
- HELIX: Hybrid Encoding with Learnable Identity and Cross-dimensional Synthesis for Time Series Imputation
- L-Drive: Beyond a Single Mapping—Latent Context Drives Time Series Forecasting
- MedMamba: Multi-View State Space Models with Adaptive Graph Learning for Medical Time Series Classification
- Nested Spatio-Temporal Time Series Forecasting
- One-Step Graph-Structured Neural Flows for Irregular Multivariate Time Series Classification
- PESD-TSF: A Period-Aware and Explicit Structured Decomposition Framework for Long-Term Time Series Forecasting
and 23 more in the interactive map.
Anomaly detection · 7 papers
- AnomSeer: Reinforcing Multimodal LLMs to Reason for Time-Series Anomaly Detection
- Cluster-Aware Causal Mixer for Online Anomaly Detection in Multivariate Time Series
- COGNOS: Universal Enhancement for Time Series Anomaly Detection via Constrained Gaussian-Noise Optimization and Smoothing
- IMPACT: Influence Modeling for Open-Set Time Series Anomaly Detection
- TimeLAVA: Learning-Agnostic Valuation for Time Series Data
- TeamWork: Multivariate Time Series Anomaly Detection via Asymmetric Role-aware Channel Modeling
- Towards Foundation Models for Zero-Shot Time Series Anomaly Detection: Leveraging Synthetic Data and Relative Context Discrepancy