Research map / Time series / Causal and generative time series
What went wrong and when? Instance-wise feature importance for time-series black-box models
Sana Tonekaboni, Shalmali Joshi, Kieran Campbell, David K. Duvenaud, Anna Goldenberg
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 FIT (temporal-shift-aware, generator-based counterfactuals) instead of gradient/attention/perturbation baselines | feature importance for time series · State data | AUROC (explanation): 0.798 vs best baseline 0.565 | Improved |
| use FIT instead of baselines | feature importance · Switch-feature data | AUROC (explanation): 0.720 vs AFO 0.590 | Improved |
| use FIT instead of baselines on a simple spike task | feature importance · Spike data | AUPRC (explanation): 0.866 vs AFO 0.932 | Mixed |
| use FIT instead of baselines on MIMIC-mortality | feature importance · MIMIC-mortality k=50 | AUROC drop: 0.133 vs FO 0.095 | Improved |
| use FIT instead of baselines on MIMIC-intervention | feature importance · MIMIC-intervention | AUROC drop: 0.050 vs GradSHAP 0.072 (k=50); 0.024 vs 0.028 (95-pc) | No clear effect |
| replace random-noise perturbations with generative-model counterfactuals (AFO to FIT) on Spike | perturbation-based importance · Spike data | AUROC (explanation): 0.942 -> 0.968 | Improved |
Closest papers by meaning
- FITS: Modeling Time Series with $10k$ Parameters (ICLR 2024 · similarity 0.83)
- TimeInf: Time Series Data Contribution via Influence Functions (ICLR 2025 · similarity 0.83)
- ShapeX: Shapelet-Driven Post Hoc Explanations for Time Series Classification Models (NeurIPS 2025 · similarity 0.83)
- A Closer Look at Transformers for Time Series Forecasting: Understanding Why They Work and Where They Struggle (ICML 2025 · similarity 0.83)
- MOMENT: A Family of Open Time-series Foundation Models (ICML 2024 · similarity 0.82)
- Feature Importance Metrics in the Presence of Missing Data (ICML 2025 · similarity 0.82)
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.
