多语言对话语音识别与说话人分离系统，显著降低错误率并获挑战赛冠亚军。
The TEA-ASLP System for Multilingual Conversational Speech Recognition and Speech Diarization in MLC-SLM 2025 Challenge
- 融合语言识别与多语言稀疏专家模型，用CTC输出作为提示提升生成质量。
- 任务一WER降至9.60%，比基线降低30.8%；任务二时间约束下WER为17.49%。
- 适用于多语言对话场景，适合关注语音识别与分离技术的开发者和研究者。
本文介绍TEA-ASLP团队提交至MLC-SLM 2025挑战赛的系统，涵盖任务一的多语言对话自动语音识别（ASR）和任务二的说话人分离语音识别。在任务一中，通过集成已知语言识别信息、多语言稀疏专家（MOE）LoRA结构，并利用CTC预测的词元作为提示，增强Ideal-LLM模型，训练数据约180k小时。任务二中，将基线英中文说话人分离模型替换为更合适的纯英语版本。该方法相较基线语音语言模型在任务一中实现30.8%的词错误率（WER）下降，最终获得9.60%的WER；任务二在时间约束下的最小排列WER为17.49%，分别获得两个任务的第一名和第二名。
原文摘要 · Abstract (English)
This paper presents the TEA-ASLP's system submitted to the MLC-SLM 2025 Challenge, addressing multilingual conversational automatic speech recognition (ASR) in Task I and speech diarization ASR in Task II. For Task I, we enhance Ideal-LLM model by integrating known language identification and a multilingual MOE LoRA structure, along with using CTC-predicted tokens as prompts to improve autoregressive generation. The model is trained on approximately 180k hours of multilingual ASR data. In Task II, we replace the baseline English-Chinese speaker diarization model with a more suitable English-only version. Our approach achieves a 30.8% reduction in word error rate (WER) compared to the baseline speech language model, resulting in a final WER of 9.60% in Task I and a time-constrained minimum-permutation WER of 17.49% in Task II, earning first and second place in the respective challenge tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。