New paper proposes predictive surprise as self-grounding concept bottleneck for interpretable time series
An arXiv paper introduces a concept bottleneck method for time-series classifiers that aims to combine accuracy, interpretability, and the ability to abstain when evidence is weak in risk-sensitive domains.
Read the full story at arXiv stat.ML ↗
Timeline · 1 report
- 2026-11-11 05:00 · arXiv stat.ML
Predictive Surprise as Self-Grounding Concept Bottleneck for Interpretable Time Series