Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval

Chen Zhao, Chenyan Xiong, Jordan Boyd-Graber, Hal Daumé III


Abstract
Complex question answering often requires finding a reasoning chain that consists of multiple evidence pieces. Current approaches incorporate the strengths of structured knowledge and unstructured text, assuming text corpora is semi-structured. Building on dense retrieval methods, we propose a new multi-step retrieval approach (BeamDR) that iteratively forms an evidence chain through beam search in dense representations. When evaluated on multi-hop question answering, BeamDR is competitive to state-of-the-art systems, without using any semi-structured information. Through query composition in dense space, BeamDR captures the implicit relationships between evidence in the reasoning chain. The code is available at https://github.com/ henryzhao5852/BeamDR.
Anthology ID:
2021.naacl-main.368
Volume:
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Month:
June
Year:
2021
Address:
Online
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4635–4641
Language:
URL:
https://aclanthology.org/2021.naacl-main.368
DOI:
10.18653/v1/2021.naacl-main.368
Bibkey:
Cite (ACL):
Chen Zhao, Chenyan Xiong, Jordan Boyd-Graber, and Hal Daumé III. 2021. Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4635–4641, Online. Association for Computational Linguistics.
Cite (Informal):
Multi-Step Reasoning Over Unstructured Text with Beam Dense Retrieval (Zhao et al., NAACL 2021)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/2021.naacl-main.368.pdf
Video:
 https://preview.aclanthology.org/ingestion-script-update/2021.naacl-main.368.mp4
Code
 henryzhao5852/BeamDR