Empirical Investigation of Neural Symbolic Reasoning Strategies

Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi, Kentaro Inui

Abstract
Neural reasoning accuracy improves when generating intermediate reasoning steps. However, the source of this improvement is yet unclear. Here, we investigate and factorize the benefit of generating intermediate steps for symbolic reasoning. Specifically, we decompose the reasoning strategy w.r.t. step granularity and chaining strategy. With a purely symbolic numerical reasoning dataset (e.g., A=1, B=3, C=A+3, C?), we found that the choice of reasoning strategies significantly affects the performance, with the gap becoming even larger as the extrapolation length becomes longer. Surprisingly, we also found that certain configurations lead to nearly perfect performance, even in the case of length extrapolation. Our results indicate the importance of further exploring effective strategies for neural reasoning models.
Anthology ID:
2023.findings-eacl.86
Volume:
Findings of the Association for Computational Linguistics: EACL 2023
Month:
May
Year:
2023
Address:
Dubrovnik, Croatia
Editors:
Andreas Vlachos, Isabelle Augenstein
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1154–1162
Language:
URL:
https://preview.aclanthology.org/build-pipeline-with-new-library/2023.findings-eacl.86/
DOI:
10.18653/v1/2023.findings-eacl.86
Bibkey:
Cite (ACL):
Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi, and Kentaro Inui. 2023. Empirical Investigation of Neural Symbolic Reasoning Strategies. In Findings of the Association for Computational Linguistics: EACL 2023, pages 1154–1162, Dubrovnik, Croatia. Association for Computational Linguistics.
Cite (Informal):
Empirical Investigation of Neural Symbolic Reasoning Strategies (Aoki et al., Findings 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/build-pipeline-with-new-library/2023.findings-eacl.86.pdf
Software:
 2023.findings-eacl.86.software.zip
Video:
 https://preview.aclanthology.org/build-pipeline-with-new-library/2023.findings-eacl.86.mp4
PDF Search Software Video Fix metadata