A masked contrastive learning framework for learning meaningful fine-grained representations with coarse-labeled dataset. read more of MaskCon: Masked Contrastive Learning for Coarse-Labeled Dataset

A self-supervised learning method aiming to alleviate the inherent false-negative problem in contrastive learning framework. read more of Adaptive Soft Contrastive Learning

A robust and efficient training framework tackling with dataset with noisy labels. read more of SSR: An Efficient and Robust Framework for Learning with Unknown Label Noise

A self-supervised pre-training method with focus on alleviating class collision problem using a cross-context learning scheme. read more of Self-Supervised Representation Learning with Cross-Context Learning between Global and Hypercolumn Features

A CLIP-based visual-language model called DFER-CLIP for in-the-wild dynamic facial expression of emotion recognition. read more of Prompting Visual-Language Models for Dynamic Facial Expression Recognition

Leverages the association between parts of speech and specific visual modes of variation to better separate representations of style from content in the CLIP representations read more of Parts of Speech–Grounded Subspaces in Vision-Language Models

Localized image editing through joint factorization of parts of appearances in pre-trained GANs. read more of PandA: Unsupervised Learning of Parts and Appearances in the Feature Maps of GANs

A method for controlling diversity between clusterings in deep clustering frameworks. read more of DivClust: Controlling Diversity in Deep Clustering

Learning Visual Explanations for DCNN-Based Image Classifiers Using an Attention Mechanism read more of L-CAM