RECTIFYING BIAS IN CLIP

Summary from the TLDR AI newsletter

Biases in CLIP can hinder its performance in unsupervised semantic segmentation, a task where images lack annotations. This paper proposes a method that explicitly models and corrects these biases.