Progressive Evidence Refinement for Open-domain Multimodal Retrieval Question Answering

TL;DR

This paper introduces a progressive evidence refinement framework with semi-supervised contrastive learning and multi-turn retrieval for open-domain multimodal question answering, significantly improving evidence selection and answer accuracy.

Contribution

It proposes a novel two-stage evidence refinement and multi-turn retrieval framework with contrastive learning to better utilize fine-grained multimodal evidence in QA tasks.

Findings

Achieved state-of-the-art results on WebQA and MultimodelQA benchmarks.

Demonstrated improved evidence selection and answer accuracy.

Validated effectiveness of the multi-turn retrieval strategy.

Abstract

Pre-trained multimodal models have achieved significant success in retrieval-based question answering. However, current multimodal retrieval question-answering models face two main challenges. Firstly, utilizing compressed evidence features as input to the model results in the loss of fine-grained information within the evidence. Secondly, a gap exists between the feature extraction of evidence and the question, which hinders the model from effectively extracting critical features from the evidence based on the given question. We propose a two-stage framework for evidence retrieval and question-answering to alleviate these issues. First and foremost, we propose a progressive evidence refinement strategy for selecting crucial evidence. This strategy employs an iterative evidence retrieval approach to uncover the logical sequence among the evidence pieces. It incorporates two rounds of filtering to optimize the solution space, thus further ensuring temporal efficiency. Subsequently, we introduce a semi-supervised contrastive learning training strategy based on negative samples to expand the scope of the question domain, allowing for a more thorough exploration of latent knowledge within known samples. Finally, in order to mitigate the loss of fine-grained information, we devise a multi-turn retrieval and question-answering strategy to handle multimodal inputs. This strategy involves incorporating multimodal evidence directly into the model as part of the historical dialogue and question. Meanwhile, we leverage a cross-modal attention mechanism to capture the underlying connections between the evidence and the question, and the answer is generated through a decoding generation approach. We validate the model's effectiveness through extensive experiments, achieving outstanding performance on WebQA and MultimodelQA benchmark tests.