[Adapt] [ADAPT SEMINAR] Can Hierachical Structure Improve Contrastive Learning? HCSC!

[Xiujie Song]

Hi Adapters,

Can hierachical structure improve contrastive learning? The paper, HCSC: Hierarchical Contrastive Selective Coding, will give you an answer!

Hierarchical semantic structures naturally exist in an image dataset, in which several semantically relevant image clusters can be further integrated into a larger cluster with coarser-grained semantics. Capturing such structures with image representations can greatly benefit the semantic understanding on various downstream tasks. Existing contrastive representation learning methods lack such an important model capability. 

In addition, the negative pairs used in these methods are not guaranteed to be semantically distinct, which could further hamper the structural
correctness of learned image representations. To tackle these limitations, some researchers propose a novel contrastive learning framework called Hierarchical Contrastive Selective Coding (HCSC).

In this framework, a set of hierarchical prototypes are constructed and also dynamically updated to represent the hierarchical semantic structures underlying the data in the latent space.

Hope you find it interesting and helpful!

Time: Wed 4:00pm
Venue: Tencent meeting 309-570-966

Best wishes,
Xiujie.