Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Instance Causal Representation Learning for Instance Label Prediction and Out-of-Distribution Generalization
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 CausalMIL instead of MIVAE | Instance-label prediction on MNIST Bags · MNIST Bags | F-score: 0.901±.035 -> 0.966±.018 | Helped |
| Use CausalMIL instead of MIVAE | Instance-label prediction on FashionMNIST Bags · FashionMNIST Bags | AUC-PR: 0.733±.271 -> 0.822±.168 | Helped |
| Use CausalMIL instead of MIVAE | Instance-label prediction on KuzushijiMNIST Bags · KuzushijiMNIST Bags | AUC-PR: 0.838±.102 -> 0.934±.075 | Helped |
| Use CausalMIL instead of MIVAE | Instance-label prediction on Colon Cancer · Colon Cancer | AUC-PR: 0.747±.032 -> 0.878±.046 | Helped |
| Use CausalMIL instead of iCaRL | OOD prediction on ColoredMNIST · ColoredMNIST test | accuracy: 0.688±.007 -> 0.892±.002 | Helped |
| Use CausalMIL instead of iCaRL | OOD prediction on ColoredFashionMNIST · ColoredFashionMNIST test | accuracy: 0.736±.006 -> 0.816±.015 | Helped |
| Remove the bag-conditional KL term (ablation) | CausalMIL representation learning (miAE) · MNIST bags, qualitative reconstructions | representation quality | Hurt |
| Remove targeted reconstruction (ablation) | CausalMIL representation learning (miVAE) · MNIST bags, qualitative reconstructions | representation quality | Hurt |
| Replace CausalMIL with an unconditional-prior MIVAE (ablation) | Representation learning for positive digit 9 · MNIST bags, qualitative reconstructions | representation quality | Hurt |
| Increase bag size (ablation) | MIL instance-label prediction methods · FashionMNIST bags | AUC-PR | Hurt |
| Increase witness rate (ablation) | Instance-independent methods such as mi-Net and Attn-MIL · FashionMNIST bags | AUC-PR | Mixed |
Closest papers by meaning
- Partial Label Causal Representation Learning for Instance-Dependent Supervision and Domain Generalization (AAAI 2025 · similarity 0.90)
- Reduction-based Pseudo-label Generation for Instance-dependent Partial Label Learning (NeurIPS 2025 · similarity 0.87)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.86)
- Distilling Reliable Knowledge for Instance-Dependent Partial Label Learning (AAAI 2024 · similarity 0.86)
- Forming Auxiliary High-confident Instance-level Loss to Promote Learning from Label Proportions (CVPR 2025 · similarity 0.86)
- Fast Multi-Instance Partial-Label Learning (AAAI 2025 · similarity 0.85)
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.