Projects per year
Abstract
In Neural Machine Translation (NMT), models will sometimes generate repetitive or fluent output that is not grounded in the source sentence. This phenomenon is known as hallucination and is a problem even in large-scale multilingual translation models. We propose to use Contrastive Decoding, an algorithm developed to improve generation from unconditional language models, to mitigate hallucinations in NMT. Specifically, we maximise the log-likelihood difference between a model and the same model with reduced contribution from the encoder outputs. Additionally, we propose an alternative implementation of Contrastive Decoding that dynamically weights the difference based on the maximum probability in the output distribution to reduce the effect of CD when the model is confident of its prediction. We evaluate our methods using the Small (418M) and Medium (1.2B) M2M models across 21 low and medium-resource language pairs. Our results show a 14.6 ± 0.5 and 11.0 ± 0.6 maximal increase in the mean COMET scores for the Small and Medium models on those sentences for which the M2M models initially generate a hallucination., respectively.
Original language | English |
---|---|
Title of host publication | Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics |
Editors | Yvette Graham, Matthew Purver |
Publisher | Association for Computational Linguistics |
Pages | 2526–2539 |
Number of pages | 14 |
Volume | 1 |
ISBN (Electronic) | 9798891760882 |
Publication status | Published - 22 Mar 2024 |
Event | The 18th Conference of the European Chapter of the Association for Computational Linguistics - St. Julian’s, Malta Duration: 17 Mar 2024 → 22 Mar 2024 Conference number: 18 https://2024.eacl.org/ |
Publication series
Name | Proceedings of the EACL Conference |
---|---|
Publisher | Association for Computational Linguistics |
ISSN (Print) | 1525-2450 |
Conference
Conference | The 18th Conference of the European Chapter of the Association for Computational Linguistics |
---|---|
Abbreviated title | EACL 2024 |
Country/Territory | Malta |
City | St. Julian’s |
Period | 17/03/24 → 22/03/24 |
Internet address |
Fingerprint
Dive into the research topics of 'Contrastive decoding reduces hallucinations in large multilingual machine translation models'. Together they form a unique fingerprint.Projects
- 1 Active