Dec 10, 2023
Speaker · 0 followers
Speaker · 0 followers
Existing contrastive learning methods rely on pairwise sample contrast z_x^⊤ z_x' to learn data representations, but the learned features often lack clear interpretability from a human perspective. Theoretically, it lacks feature identifiability and different initialization may lead to totally different features.In this paper, we study a new method named tri-contrastive learning (triCL) that involves a 3-factor contrast in the form of z_x^⊤ S z_x', where S= diag(s_1,...,s_k) is a learnable diagonal matrix that automatically captures the importance of each feature. We show that by this simple extension, triCL can not only obtain identifiable features that eliminate randomness, but also obtain more interpretable features that are ordered according to the importance matrix S. We show that features with high importance have nice interpretability by capturing common classwise features, and obtain superior performance when evaluated for image retrieval using a few features. The proposed triCL objective is general and can be applied to different contrastive learning methods like SimCLR and CLIP. We believe that it is a better alternative to existing 2-factor contrastive learning by improving its identifiability and interpretability with minimal overhead.Existing contrastive learning methods rely on pairwise sample contrast z_x^⊤ z_x' to learn data representations, but the learned features often lack clear interpretability from a human perspective. Theoretically, it lacks feature identifiability and different initialization may lead to totally different features.In this paper, we study a new method named tri-contrastive learning (triCL) that involves a 3-factor contrast in the form of z_x^⊤ S z_x', where S= diag(s_1,...,s_k) is a learnable diago…
Account · 617 followers
Professional recording and live streaming, delivered globally.
Presentations on similar topic, category or speaker
Han Hu, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Puheng Li, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Kate Sanders, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Jialin Chen, …
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%
Total of 0 viewers voted for saving the presentation to eternal vault which is 0.0%