MIX : a Multi-task Learning Approach to Solve Open-Domain Question Answering

TL;DR

MIX is a multi-task deep learning system designed for open-domain question answering, combining retrieval, scoring, and extraction components to achieve competitive performance with improved efficiency.

Contribution

The paper introduces a multi-task learning framework that parallelizes key components in open-domain QA, simplifying the system while maintaining state-of-the-art results.

Findings

Achieves performance comparable to state-of-the-art on squad-open.

Improves computational efficiency through multi-task parallelization.

Simplifies the QA pipeline without sacrificing accuracy.

Abstract

This paper introduces MIX, a multi-task deep learning approach to solve open-ended question-answering. First, we design our system as a multi-stage pipeline of 3 building blocks: a BM25-based Retriever to reduce the search space, a RoBERTa-based Scorer, and an Extractor to rank retrieved paragraphs and extract relevant text spans, respectively. Eventually, we further improve the computational efficiency of our system to deal with the scalability challenge: thanks to multi-task learning, we parallelize the close tasks solved by the Scorer and the Extractor. Our system is on par with state-of-the-art performances on the squad-open benchmark while being simpler conceptually.