Zhiyuan Chang

New approach improves testing for visual grounding models using combined text and image analysis.

A new method enhances the testing of visual entailment systems by focusing on object relationships.

Learn how external knowledge improves language model accuracy and reliability.