Yueqian Wang

New benchmark assesses how video-language models handle inaccuracies effectively.

A new model allows real-time interaction with videos, enhancing understanding and engagement.

Research reveals how we can make machines understand complex dialogues.