Repository logo
 

Retrieval Augmented Visual Question Answering with Outside Knowledge

Published version
Peer-reviewed

Type

Conference Object

Change log

Authors

Byrne, B 

Abstract

Outside-Knowledge Visual Question Answering (OK-VQA) is a challenging VQA task that requires retrieval of external knowledge to answer questions about images. Recent OK-VQA systems use Dense Passage Retrieval (DPR) to retrieve documents from external knowledge bases, such as Wikipedia, but with DPR trained separately from answer generation, introducing a potential limit on the overall system performance. Instead, we propose a joint training scheme which includes differentiable DPR integrated with answer generation so that the system can be trained in an end-to-end fashion. Our experiments show that our scheme outperforms recent OK-VQA systems with strong DPR for retrieval. We also introduce new diagnostic metrics to analyze how retrieval and generation interact. The strong retrieval ability of our model significantly reduces the number of retrieved documents needed in training, yielding significant benefits in answer quality and computation required for training.

Description

Keywords

Journal Title

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, EMNLP 2022

Conference Name

Journal ISSN

Volume Title

Publisher

Publisher DOI

Publisher URL