BUT System for the MLC-SLM Challenge

BUT System for the MLC-SLM Challenge

Stať ve sborníku mimo WoS a Scopus

We present a two-speaker automatic speech recognition (ASR) system that combines DiCoW—a diarization-conditioned variant of Whisper—with DiariZen, a diarization pipeline built on top of Pyannote. We first evaluate both systems in out-of-domain (OOD) multilingual scenarios without any fine-tuning. In this scenario, DiariZen consistently outperforms the baseline Pyannote diarization model, demonstrating strong generalization. Despite being fine-tuned on English-only data for target-speaker ASR, DiCoW retains solid multilingual performance,indicating that encoder modifications preserve Whisper’s multilingual capabilities. We then fine-tune both DiCoW and DiariZen on the MLC-SLM challenge data. The fine-tuned DiariZen continues to outperform the fine-tuned Pyannote baseline, while DiCoW sees further gains from domain adaptation. Our final system achieves a micro-average tcpWER/CER of 16.75 % and ranks second in Task 2 of the MLC-SLM challenge. Lastly, we identify several labeling inconsistencies in the training data—such as missing speech segments and incorrect silence annotations—which can hinder diarization fine-tuning. We propose simple mitigation strategies to address these issues and improve system robustness.

We present a two-speaker automatic speech recognition (ASR) system that combines DiCoW—a diarization-conditioned variant of Whisper—with DiariZen, a diarization pipeline built on top of Pyannote. We first evaluate both systems in out-of-domain (OOD) multilingual scenarios without any fine-tuning. In this scenario, DiariZen consistently outperforms the baseline Pyannote diarization model, demonstrating strong generalization. Despite being fine-tuned on English-only data for target-speaker ASR, DiCoW retains solid multilingual performance,indicating that encoder modifications preserve Whisper’s multilingual capabilities. We then fine-tune both DiCoW and DiariZen on the MLC-SLM challenge data. The fine-tuned DiariZen continues to outperform the fine-tuned Pyannote baseline, while DiCoW sees further gains from domain adaptation. Our final system achieves a micro-average tcpWER/CER of 16.75 % and ranks second in Task 2 of the MLC-SLM challenge. Lastly, we identify several labeling inconsistencies in the training data—such as missing speech segments and incorrect silence annotations—which can hinder diarization fine-tuning. We propose simple mitigation strategies to address these issues and improve system robustness.

DiCoW, Multilingual Multi-Talker ASR, DiariZen, Whisper

DiCoW, Multilingual Multi-Talker ASR, DiariZen, Whisper

Alexander Polok, Jiangyu Han, Dominik Klement, Samuele Cornell, Jan Černocký, Lukáš Burget

2026

22.08.2025

ISCA

ISCA

23

27

5

https://www.isca-archive.org/mlcslm_2025/polok25_mlcslm.pdf

polok_mlcslm_interspeech-2025_satelite workshop