Twist Decoding: Diverse Generators Guide Each Other

Jungo Kasai, Keisuke Sakaguchi, Ronan Le Bras, Hao Peng, Ximing Lu, Dragomir Radev, Yejin Choi, Noah A. Smith


Abstract
Many language generation models are now available for a wide range of generation tasks, including machine translation and summarization. Combining such diverse models may lead to further progress, but ensembling generation models is challenging during inference: conventional ensembling methods (e.g., shallow fusion) require that the models share vocabulary/tokenization schemes. We introduce Twist decoding, a simple and general text generation algorithm that benefits from diverse models at inference time. Our method does not assume the vocabulary, tokenization or even generation order is shared. Our extensive evaluations on machine translation and scientific paper summarization demonstrate that Twist decoding substantially outperforms each model decoded in isolation over various scenarios, including cases where domain-specific and general-purpose models are both available. Twist decoding also consistently outperforms the popular reranking heuristic where output candidates from one model are rescored by another. We hope that our work will encourage researchers and practitioners to examine generation models collectively, not just independently, and to seek out models with complementary strengths to the currently available models.
Anthology ID:
2022.emnlp-main.326
Volume:
Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4909–4923
Language:
URL:
https://aclanthology.org/2022.emnlp-main.326
DOI:
10.18653/v1/2022.emnlp-main.326
Bibkey:
Cite (ACL):
Jungo Kasai, Keisuke Sakaguchi, Ronan Le Bras, Hao Peng, Ximing Lu, Dragomir Radev, Yejin Choi, and Noah A. Smith. 2022. Twist Decoding: Diverse Generators Guide Each Other. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 4909–4923, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Twist Decoding: Diverse Generators Guide Each Other (Kasai et al., EMNLP 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-1/2022.emnlp-main.326.pdf