Unified Language Representation for Question Answering over Text, Tables, and Images

TL;DR

This paper introduces Solar, a framework that converts images and tables into unified language representations, enabling effective question answering across multiple data modalities with improved accuracy over existing methods.

Contribution

The paper proposes transforming visual and tabular data into language form, simplifying multi-modal QA into a textual problem, leveraging pre-trained language models for better performance.

Findings

Solar outperforms existing methods by 10.6-32.3 points on multiple datasets.

Achieves top results on WebQA leaderboard.

Effective cross-modal reasoning with unified language representations.

Abstract

When trying to answer complex questions, people often rely on multiple sources of information, such as visual, textual, and tabular data. Previous approaches to this problem have focused on designing input features or model structure in the multi-modal space, which is inflexible for cross-modal reasoning or data-efficient training. In this paper, we call for an alternative paradigm, which transforms the images and tables into unified language representations, so that we can simplify the task into a simpler textual QA problem that can be solved using three steps: retrieval, ranking, and generation, all within a language space. This idea takes advantage of the power of pre-trained language models and is implemented in a framework called Solar. Our experimental results show that Solar outperforms all existing methods by 10.6-32.3 pts on two datasets, MultimodalQA and MMCoQA, across ten different metrics. Additionally, Solar achieves the best performance on the WebQA leaderboard