Multi-Instance Causal Representation Learning for Instance Label Prediction and Out-of-Distribution Generalization
Nov 28, 2022
Speakers
Weijia Zhang
Speaker · 0 followers
Xuanhui Zhang
Speaker · 0 followers
Han-Wen Deng
Speaker · 0 followers
Min-Ling Zhang
Speaker · 0 followers
About
Multi-instance learning (MIL) deals with objects represented as bags of instances and can predict instance labels using only bag-level supervision. However, significant performance gaps exist between instance-level MIL algorithms and supervised learners since the instance labels are unavailable in MIL. Most existing MIL algorithms tackle the problem by treating multi-instance bags as harmful ambiguities and predicting instance labels by reducing their supervision inexactness. This work studies MIL from a new perspective by considering bags as important auxiliary information that helps identify invariant causal representations from bag-level weak supervision. We propose the TargetedMIL algorithm, which not only excels at instance label prediction but also is robust to distribution change by synergistically integrating MIL with identifiable variational autoencoder based on a practical and general assumption: the prior distribution over the instance latent representations belongs to the non-factorized exponential family conditioning on the bags. Experiments on synthetic and real-world datasets demonstrate that our approach significantly outperforms various baselines on both instance label prediction and out-of-distribution generalization tasks.Multi-instance learning (MIL) deals with objects represented as bags of instances and can predict instance labels using only bag-level supervision. However, significant performance gaps exist between instance-level MIL algorithms and supervised learners since the instance labels are unavailable in MIL. Most existing MIL algorithms tackle the problem by treating multi-instance bags as harmful ambiguities and predicting instance labels by reducing their supervision inexactness. This work studies M…
Organizer
NeurIPS 2022
Account · 953 followers
Like the format? Trust SlidesLive to capture your next event!
Professional recording and live streaming, delivered globally.
Sharing
Recommended Videos
Presentations on similar topic, category or speaker
Counterfactual Temporal Point Processes
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Monitoring of Perception Systems: Deterministic, Probabilistic, and Learning-based Fault Detection and Identification
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Closing Remarks
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Featured Papers Panels 3C
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
MSDS: A Large-Scale Chinese Signature and Token Digit String Dataset for Handwriting Verification
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Learning Generalist Agents
Watch later
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%