Research map / Time series / Causal and generative time series
TimeSeg: An Information-Theoretic Segment-Wise Explainer for Time-Series Predictions
Hwijin Kim, Jaeho Kim, Changhee Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use TimeSeg instead of TimeX++ on MIT-ECG | explainer · MIT-ECG | F1: 0.593 -> 0.739 | Improved |
| use TimeSeg instead of LIMESegment on SeqComb-UV | explainer · SeqComb-UV | F1: 0.430 -> 0.645 | Improved |
| use TimeSeg instead of TimeX++ on FreqShapes-V | explainer · FreqShapes-V | F1: 0.799 -> 0.722 | Hurt |
| use TimeSeg instead of TimeX++ on Wafer | explainer · Wafer mean substitution | sufficiency drop: 52.39 -> 0.19 | Improved |
| use TimeSeg segments instead of point-wise methods | explainer · SeqComb-UV | contiguity: 0.215 -> 0.017 | Improved |
| raise lambda from 0.1 to 0.9 (ablation) | TimeSeg · MIT-ECG | sparsity: 0.218 -> 0.098 | Improved |
| raise lambda from 0.1 to 0.9 (ablation) | TimeSeg · MIT-ECG | sufficiency: 0.558 -> 4.728 | Hurt |
| lambda 0.3 versus 0.1 (ablation) | TimeSeg · MIT-ECG | F1: 0.702 -> 0.739 | Improved |
| allow multiple segments, Kmax 1 to 3 (ablation) | TimeSeg · SeqComb-UV | F1: 0.499 -> 0.652 | Improved |
| increase Kmax from 3 to 7 (ablation) | TimeSeg · SeqComb-UV | F1: 0.652 -> 0.642 | No clear effect |
Closest papers by meaning
- Explain Temporal Black-Box Models via Functional Decomposition (ICML 2024 · similarity 0.89)
- Segment, Shuffle, and Stitch: A Simple Layer for Improving Time-Series Representations (NeurIPS 2024 · similarity 0.88)
- Optimal Information Retention for Time-Series Explanations (ICML 2025 · similarity 0.87)
- InteDisUX: Intepretation-Guided Discriminative User-Centric Explanation for Time Series (AAAI 2025 · similarity 0.87)
- Large Pre-trained time series models for cross-domain Time series analysis tasks (NeurIPS 2024 · similarity 0.87)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
