Triple-Encoders: Representations That Fire Together, Wire Together

Justus-Jonas Erker, Florian Mai, Nils Reimers, Gerasimos Spanakis, Iryna Gurevych


Abstract
Search-based dialog models typically re-encode the dialog history at every turn, incurring high cost.Curved Contrastive Learning, a representation learning method that encodes relative distances between utterances into the embedding space via a bi-encoder, has recently shown promising results for dialog modeling at far superior efficiency.While high efficiency is achieved through independently encoding utterances, this ignores the importance of contextualization. To overcome this issue, this study introduces triple-encoders, which efficiently compute distributed utterance mixtures from these independently encoded utterances through a novel hebbian inspired co-occurrence learning objective in a self-organizing manner, without using any weights, i.e., merely through local interactions. Empirically, we find that triple-encoders lead to a substantial improvement over bi-encoders, and even to better zero-shot generalization than single-vector representation models without requiring re-encoding. Our code (https://github.com/UKPLab/acl2024-triple-encoders) and model (https://huggingface.co/UKPLab/triple-encoders-dailydialog) are publicly available.
Anthology ID:
2024.acl-long.290
Volume:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5317–5332
Language:
URL:
https://aclanthology.org/2024.acl-long.290
DOI:
10.18653/v1/2024.acl-long.290
Bibkey:
Cite (ACL):
Justus-Jonas Erker, Florian Mai, Nils Reimers, Gerasimos Spanakis, and Iryna Gurevych. 2024. Triple-Encoders: Representations That Fire Together, Wire Together. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5317–5332, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
Triple-Encoders: Representations That Fire Together, Wire Together (Erker et al., ACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.acl-long.290.pdf