Blending LLMs into Cascaded Speech Translation: KIT’s Offline Speech Translation System for IWSLT 2024

Sai Koneru, Thai Binh Nguyen, Ngoc-Quan Pham, Danni Liu, Zhaolin Li, Alexander Waibel, Jan Niehues


Abstract
Large Language Models (LLMs) are currently under exploration for various tasks, including Automatic Speech Recognition (ASR), Machine Translation (MT), and even End-to-End Speech Translation (ST). In this paper, we present KIT’s offline submission in the constrained + LLM track by incorporating recently proposed techniques that can be added to any cascaded speech translation. Specifically, we integrate Mistral-7B into our system to enhance it in two ways. Firstly, we refine the ASR outputs by utilizing the N-best lists generated by our system and fine-tuning the LLM to predict the transcript accurately. Secondly, we refine the MT outputs at the document level by fine-tuning the LLM, leveraging both ASR and MT predictions to improve translation quality. We find that integrating the LLM into the ASR and MT systems results in an absolute improvement of 0.3% in Word Error Rate and 0.65% in COMET for tst2019 test set. In challenging test sets with overlapping speakers and background noise, we find that integrating LLM is not beneficial due to poor ASR performance. Here, we use ASR with chunked long-form decoding to improve context usage that may be unavailable when transcribing with Voice Activity Detection segmentation alone.
Anthology ID:
2024.iwslt-1.24
Volume:
Proceedings of the 21st International Conference on Spoken Language Translation (IWSLT 2024)
Month:
August
Year:
2024
Address:
Bangkok, Thailand (in-person and online)
Editors:
Elizabeth Salesky, Marcello Federico, Marine Carpuat
Venue:
IWSLT
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
183–191
Language:
URL:
https://aclanthology.org/2024.iwslt-1.24
DOI:
10.18653/v1/2024.iwslt-1.24
Bibkey:
Cite (ACL):
Sai Koneru, Thai Binh Nguyen, Ngoc-Quan Pham, Danni Liu, Zhaolin Li, Alexander Waibel, and Jan Niehues. 2024. Blending LLMs into Cascaded Speech Translation: KIT’s Offline Speech Translation System for IWSLT 2024. In Proceedings of the 21st International Conference on Spoken Language Translation (IWSLT 2024), pages 183–191, Bangkok, Thailand (in-person and online). Association for Computational Linguistics.
Cite (Informal):
Blending LLMs into Cascaded Speech Translation: KIT’s Offline Speech Translation System for IWSLT 2024 (Koneru et al., IWSLT 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.iwslt-1.24.pdf