Research map / Label-efficient and robust learning / Recognition and long-tail
Model Agnostic Interpretability for Multiple Instance Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 MILLI instead of the strongest listed baseline | Four MIL models · 4-MNIST-Bags | NDCG@n: 0.833 -> 0.942 | Helped |
| Use MILLI instead of GuidedSHAP | Four MIL models · CRC | NDCG@n: 0.827 -> 0.786 | Hurt |
| Use MILLI instead of Combined | Four MIL models · SIVAL | NDCG@n: 0.805 -> 0.808 | Helped |
| Use MILLI instead of RandomLIME | Four MIL models · SIVAL | AOPC-R: 0.299 -> 0.284 | Hurt |
| Add single-instance attribution (ablation) | Independent-instance attribution methods · CRC | NDCG@n: 0.724 -> 0.823 | Helped |
| Add one-removed attribution (ablation) | Independent-instance attribution methods · SIVAL | NDCG@n: 0.793 -> 0.782 | Hurt |
| Combine Single and One Removed attributions (ablation) | Independent-instance attribution methods · SIVAL | NDCG@n: 0.793 -> 0.805 | Helped |
| Combine Single and One Removed attributions (ablation) | Independent-instance attribution methods · 4-MNIST-Bags | NDCG@n: 0.746 -> 0.792 | Helped |
| Use single-instance attribution rather than one-removed attribution (ablation) | Independent-instance attribution methods · CRC | NDCG@n: 0.724 -> 0.823 | Helped |
| Sample coalitions with MILLI to capture instance interactions | Four MIL models · 4-MNIST-Bags | NDCG@n: 0.833 -> 0.942 | Helped |
| Use coalitional local-surrogate methods on independent-instance CRC data | CRC interpretability · CRC | NDCG@n: GuidedSHAP 0.827; MILLI 0.786 | Mixed |
Closest papers by meaning
- An Additive Instance-Wise Approach to Multi-class Model Interpretation (ICLR 2023 · similarity 0.89)
- Reproducibility in Multiple Instance Learning: A Case For Algorithmic Unit Tests (NeurIPS 2023 · similarity 0.87)
- Are Multiple Instance Learning Algorithms Learnable for Instances? (NeurIPS 2024 · similarity 0.87)
- Multi-Instance Causal Representation Learning for Instance Label Prediction and Out-of-Distribution Generalization (NeurIPS 2022 · similarity 0.86)
- Inherently Interpretable Time Series Classification via Multiple Instance Learning (ICLR 2024 · similarity 0.85)
- What makes an Ensemble (Un) Interpretable? (ICML 2025 · similarity 0.83)
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.