Multi-Label Transfer Learning for Multi-Relational Semantic Similarity

Li Zhang, Steven Wilson, Rada Mihalcea


Abstract
Multi-relational semantic similarity datasets define the semantic relations between two short texts in multiple ways, e.g., similarity, relatedness, and so on. Yet, all the systems to date designed to capture such relations target one relation at a time. We propose a multi-label transfer learning approach based on LSTM to make predictions for several relations simultaneously and aggregate the losses to update the parameters. This multi-label regression approach jointly learns the information provided by the multiple relations, rather than treating them as separate tasks. Not only does this approach outperform the single-task approach and the traditional multi-task learning approach, but it also achieves state-of-the-art performance on all but one relation of the Human Activity Phrase dataset.
Anthology ID:
S19-1005
Volume:
Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM 2019)
Month:
June
Year:
2019
Address:
Minneapolis, Minnesota
Editors:
Rada Mihalcea, Ekaterina Shutova, Lun-Wei Ku, Kilian Evang, Soujanya Poria
Venue:
*SEM
SIGs:
SIGSEM | SIGLEX
Publisher:
Association for Computational Linguistics
Note:
Pages:
44–50
Language:
URL:
https://aclanthology.org/S19-1005
DOI:
10.18653/v1/S19-1005
Bibkey:
Cite (ACL):
Li Zhang, Steven Wilson, and Rada Mihalcea. 2019. Multi-Label Transfer Learning for Multi-Relational Semantic Similarity. In Proceedings of the Eighth Joint Conference on Lexical and Computational Semantics (*SEM 2019), pages 44–50, Minneapolis, Minnesota. Association for Computational Linguistics.
Cite (Informal):
Multi-Label Transfer Learning for Multi-Relational Semantic Similarity (Zhang et al., *SEM 2019)
Copy Citation:
PDF:
https://aclanthology.org/S19-1005.pdf
Presentation:
 S19-1005.Presentation.pdf
Data
SICK