研究人员开发了一种新颖的方法，通过生成合成对话数据来增强低资源语言的自动语音识别（ASR）训练。该流程使用LLMs创建对话，将说话者属性映射到TTS语音配置文件，并组装模拟对话。在匈牙利BEA-Dialogue基准上的评估表明，这种合成数据显著提高了ASR性能，甚至优于在更大真实数据集上训练的模型。 AI
影响 通过LLMs和TTS进行合成数据生成，为改善低资源语言的ASR提供了可扩展的解决方案。
排序理由 该集群包含一篇详细介绍ASR模型训练新方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.03957v1 Announce Type: cross Abstract: Conversational ASR for lower-resource languages and niche domains is limited by the scarcity of domain-matched multi-speaker training data. We propose an augmentation pipeline that generates scenario-level dialogues with participa…
Conversational ASR for lower-resource languages and niche domains is limited by the scarcity of domain-matched multi-speaker training data. We propose an augmentation pipeline that generates scenario-level dialogues with participant metadata, maps speaker attributes to TTS voice …
Conversational ASR for lower-resource languages and niche domains is limited by the scarcity of domain-matched multi-speaker training data. We propose an augmentation pipeline that generates scenario-level dialogues with participant metadata, maps speaker attributes to TTS voice …