On the Generalization Ability of Retrieval-Enhanced Transformers

Tobias Norlund, Ehsan Doostmohammadi, Richard Johansson, Marco Kuhlmann


Abstract
Recent work on the Retrieval-Enhanced Transformer (RETRO) model has shown impressive results: off-loading memory from trainable weights to a retrieval database can significantly improve language modeling and match the performance of non-retrieval models that are an order of magnitude larger in size. It has been suggested that at least some of this performance gain is due to non-trivial generalization based on both model weights and retrieval. In this paper, we try to better understand the relative contributions of these two components. We find that the performance gains from retrieval to a very large extent originate from overlapping tokens between the database and the test data, suggesting less of non-trivial generalization than previously assumed. More generally, our results point to the challenges of evaluating the generalization of retrieval-augmented language models such as RETRO, as even limited token overlap may significantly decrease test-time loss. We release our code and model at https://github.com/TobiasNorlund/retro
Anthology ID:
2023.findings-eacl.109
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:
1485–1493
Language:
URL:
https://aclanthology.org/2023.findings-eacl.109
DOI:
10.18653/v1/2023.findings-eacl.109
Bibkey:
Cite (ACL):
Tobias Norlund, Ehsan Doostmohammadi, Richard Johansson, and Marco Kuhlmann. 2023. On the Generalization Ability of Retrieval-Enhanced Transformers. In Findings of the Association for Computational Linguistics: EACL 2023, pages 1485–1493, Dubrovnik, Croatia. Association for Computational Linguistics.
Cite (Informal):
On the Generalization Ability of Retrieval-Enhanced Transformers (Norlund et al., Findings 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.findings-eacl.109.pdf
Video:
 https://aclanthology.org/2023.findings-eacl.109.mp4