ICLR 2025 accepted papers on Time series

Causal and generative time series · 26 papers

• ZAPBench: A Benchmark for Whole-Brain Activity Prediction in Zebrafish (2 citations)
• Continuous Ensemble Weather Forecasting with Diffusion models (1 citation)
• HR-Extreme: A High-Resolution Dataset for Extreme Weather Forecasting (1 citation)
• Kernel-based Optimally Weighted Conformal Time-Series Prediction (1 citation)
• Learning Interpretable Hierarchical Dynamical Systems Models from Time Series Data (1 citation)
• Oscillatory State-Space Models (1 citation)
• SigDiffusions: Score-Based Diffusion Models for Time Series via Log-Signature Embeddings (1 citation)
• Amortized Control of Continuous State Space Feynman-Kac Model for Irregular Time Series
• Compositional simulation-based inference for time series
• Dynamical Diffusion: Learning Temporal Dynamics with Diffusion Models
• Error-quantified Conformal Inference for Time Series
• Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting

and 14 more in the interactive map.

Foundation models for time series · 19 papers

• Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (11 citations)
• Specialized Foundation Models Struggle to Beat Supervised Baselines (6 citations)
• Towards Neural Scaling Laws for Time Series Foundation Models (2 citations)
• Can LLMs Understand Time Series Anomalies? (1 citation)
• Fourier Head: Helping Large Language Models Learn Complex Probability Distributions (1 citation)
• In-context Time Series Predictor (1 citation)
• Learning under Temporal Label Noise (1 citation)
• Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model (1 citation)
• Zero-shot forecasting of chaotic systems (1 citation)
• AstroCompress: A benchmark dataset for multi-purpose compression of astronomical data
• Density estimation with LLMs: a geometric investigation of in-context learning trajectories
• Differentiable Rule Induction from Raw Sequence Inputs

and 7 more in the interactive map.

Long-term forecasting · 14 papers

• FreDF: Learning to Forecast in the Frequency Domain (22 citations)
• Timer-XL: Long-Context Transformers for Unified Time Series Forecasting (6 citations)
• TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis (4 citations)
• TimeKAN: KAN-based Frequency Decomposition Learning Architecture for Long-term Time Series Forecasting (3 citations)
• CoMRes: Semi-Supervised Time Series Forecasting Utilizing Consensus Promotion of Multi-Resolution
• FACTS: A Factored State-Space Framework for World Modelling
• Locally Connected Echo State Networks for Time Series Forecasting
• S4M: S4 for multivariate time series forecasting with Missing values
• TVNet: A Novel Time Series Analysis Method Based on Dynamic Convolution and 3D-Variation
• Fast and Slow Streams for Online Time Series Forecasting Without Information Leakage
• Going Beyond Static: Understanding Shifts with Time-Series Attribution
• Investigating Pattern Neurons in Urban Time Series Forecasting

and 2 more in the interactive map.

Anomaly detection · 5 papers

• CATCH: Channel-Aware Multivariate Time Series Anomaly Detection via Frequency Patching (2 citations)
• Towards a General Time Series Anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders
• Learn hybrid prototypes for multivariate time series anomaly detection
• Multi-Resolution Decomposable Diffusion Model for Non-Stationary Time Series Anomaly Detection
• TimeInf: Time Series Data Contribution via Influence Functions