Open Domain Question Answering over Tables via Dense Retrieval

TL;DR

This paper introduces a novel dense retrieval method for open-domain question answering over tables, significantly improving retrieval and QA performance by designing a tabular context-aware retriever and pre-training it effectively.

Contribution

It is the first to address open-domain QA over tables with a specialized retriever and pre-training, enhancing retrieval and answer accuracy.

Findings

Retrieval recall improved from 72.0 to 81.1

End-to-end QA accuracy increased from 33.8 to 37.7

Effective pre-training and hard negative mining boost performance

Abstract

Recent advances in open-domain QA have led to strong models based on dense retrieval, but only focused on retrieving textual passages. In this work, we tackle open-domain QA over tables for the first time, and show that retrieval can be improved by a retriever designed to handle tabular context. We present an effective pre-training procedure for our retriever and improve retrieval quality with mined hard negatives. As relevant datasets are missing, we extract a subset of Natural Questions (Kwiatkowski et al., 2019) into a Table QA dataset. We find that our retriever improves retrieval results from 72.0 to 81.1 recall@10 and end-to-end QA results from 33.8 to 37.7 exact match, over a BERT based retriever.