<ul><li>Researchers have introduced ESCHER, a visual concept library that aims to improve visual recognition.</li><li>ESCHER utilizes a vision-language model as a critic to iteratively refine the concept library.</li><li>The approach considers interactions between concepts and their impact on downstream classifiers.</li><li>ESCHER does not require human annotations and demonstrates effectiveness in various visual classification tasks.</li></ul>

Self-Evolving Visual Concept Library using Vision-Language Critics

Discover more