
RACC, a framework that learns to compress and aggregate retrieved contexts, achieves state-of-the-art performance on knowledge-based visual question answering tasks while significantly reducing inference latency and storage requirements.


coremsg

efficient-knowledge-based-visual-question-answering-by-learning-to-compress-and-aggregate-contextual-information


Efficient Knowledge-based Visual Question Answering by Learning to Compress and Aggregate Contextual Information
