Arc Loss: Softmax with Additive Angular Margin for Answer Retrieval

Rikiya Suzuki, Sumio Fujita, Tetsuya Sakai

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Answer retrieval is a crucial step in question answering. To determine the best Q–A pair in a candidate pool, traditional approaches adopt triplet loss (i.e., pairwise ranking loss) for a meaningful distributed representation. Triplet loss is widely used to push away a negative answer from a certain question in a feature space and leads to a better understanding of the relationship between questions and answers. However, triplet loss is inefficient because it requires two steps: triplet generation and negative sampling. In this study, we propose an alternative loss function, namely, arc loss, for more efficient and effective learning than that by triplet loss. We evaluate the proposed approach on a commonly used QA dataset and demonstrate that it significantly outperforms the triplet loss baseline.

Original languageEnglish
Title of host publicationInformation Retrieval Technology - 15th Asia Information Retrieval Societies Conference, AIRS 2019, Proceedings
EditorsFu Lee Wang, Haoran Xie, Wai Lam, Aixin Sun, Lun-Wei Ku, Tianyong Hao, Wei Chen, Tak-Lam Wong, Xiaohui Tao
PublisherSpringer
Pages34-40
Number of pages7
ISBN (Print)9783030428341
DOIs
Publication statusPublished - 2020
Event15th Asia Information Retrieval Societies Conference, AIRS 2019 - Kowloon, Hong Kong
Duration: 2019 Nov 72019 Nov 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12004 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th Asia Information Retrieval Societies Conference, AIRS 2019
CountryHong Kong
CityKowloon
Period19/11/719/11/9

Keywords

  • Answer retrieval
  • Question answering
  • Representation learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'Arc Loss: Softmax with Additive Angular Margin for Answer Retrieval'. Together they form a unique fingerprint.

  • Cite this

    Suzuki, R., Fujita, S., & Sakai, T. (2020). Arc Loss: Softmax with Additive Angular Margin for Answer Retrieval. In F. L. Wang, H. Xie, W. Lam, A. Sun, L-W. Ku, T. Hao, W. Chen, T-L. Wong, & X. Tao (Eds.), Information Retrieval Technology - 15th Asia Information Retrieval Societies Conference, AIRS 2019, Proceedings (pp. 34-40). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12004 LNCS). Springer. https://doi.org/10.1007/978-3-030-42835-8_4