Research map / Time series / Causal and generative time series

TSB-SEG: A Systematic Time-Series Segmentation Benchmark

NeurIPS 2026 · Time series › Causal and generative time series · approach: Robust forecasting and evaluation

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Introduces TSB-SEG, a benchmark for time-series segmentation that unifies change point detection and state detection evaluation to address fragmentation between the statistical and ML communities.

Problem

Time-series segmentation is studied separately as change point detection (statistics) and state detection (ML/data mining), leading to fragmented literature and evaluations that do not transfer across the two communities.

Approach

A unified systematic benchmark that evaluates segmentation methods from both the CPD and SD traditions under common protocols.

Experimental setup

Benchmarks: TSB-SEG

Closest papers by meaning

• The Elephant in the Room: Towards A Reliable Time-Series Anomaly Detection Benchmark (NeurIPS 2024 · similarity 0.89)
• Toward Interpretable Evaluation Measures for Time Series Segmentation (NeurIPS 2025 · similarity 0.89)
• Detect, Explain, Interpret: An End-to-End Benchmark for Time Series Anomaly Detection, Explainability and Interpretability. (NeurIPS 2026 · similarity 0.86)
• Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection (NeurIPS 2025 · similarity 0.86)
• TimeMixer++: A General Time Series Pattern Machine for Universal Predictive Analysis (ICLR 2025 · similarity 0.85)
• Entropy Guided Dynamic Patch Segmentation for Time Series Transformers (NeurIPS 2026 · similarity 0.85)