RAMQA: A Unified Framework for Retrieval-Augmented Multi-Modal Question Answering

TL;DR

RAMQA introduces a unified framework that combines learning-to-rank and generative models to enhance multi-modal question answering by effectively integrating text and images, leading to significant performance improvements.

Contribution

It proposes a novel combination of ranking and generative techniques using LLaVA and LLaMA models for improved multi-modal QA.

Findings

Significant performance improvements on WebQA and MultiModalQA benchmarks.

Effective integration of ranking and generative models for multi-modal QA.

Demonstrated superiority over strong baseline methods.

Abstract

Multi-modal retrieval-augmented Question Answering (MRAQA), integrating text and images, has gained significant attention in information retrieval (IR) and natural language processing (NLP). Traditional ranking methods rely on small encoder-based language models, which are incompatible with modern decoder-based generative large language models (LLMs) that have advanced various NLP tasks. To bridge this gap, we propose RAMQA, a unified framework combining learning-to-rank methods with generative permutation-enhanced ranking techniques. We first train a pointwise multi-modal ranker using LLaVA as the backbone. Then, we apply instruction tuning to train a LLaMA model for re-ranking the top-k documents using an innovative autoregressive multi-task learning approach. Our generative ranking model generates re-ranked document IDs and specific answers from document candidates in various permutations. Experiments on two MRAQA benchmarks, WebQA and MultiModalQA, show significant improvements over strong baselines, highlighting the effectiveness of our approach. Code and data are available at: https://github.com/TonyBY/RAMQA