针对语音识别中语言差异导致的性能不均，提出新优化方法提升弱势语言表现。
CTC-DRO: Robust Optimization for Reducing Language Disparities in Speech Recognition
- 通过平滑分组权重更新，避免过度关注持续高误差的语言组。
- 在多语言语音识别中，最差语言错误率降低47.1%，平均错误率降32.9%。
- 适用于有长序列与语言差异问题的语音任务，计算开销小。
现代深度学习模型虽整体性能高，但在特定子群体上表现不佳。组分布鲁棒优化（group DRO）通过最小化最差组损失来缓解此问题，但当组损失无法准确反映组间性能差异时会失效。这在语音领域尤为常见，因广泛使用的连接时序分类（CTC）损失不仅随输入长度变化，还受语言和声学特性影响，导致组损失产生虚假差异。本文提出CTC-DRO，通过平滑组权重更新以避免对持续高损失组过度强调，并采用按输入长度匹配的批处理方式缓解CTC的缩放问题。我们在ML-SUPERB 2.0基准的五个语言集上评估了多语言自动语音识别（ASR）任务，结果表明CTC-DRO始终优于group DRO和基于CTC的基线模型，最差语言错误率降低最高达47.1%，平均错误率降低最高达32.9%。该方法可低开销应用于ASR，且对存在类似挑战的其他领域具有推广潜力。
原文摘要 · Abstract (English)
Modern deep learning models often achieve high overall performance, but consistently fail on specific subgroups. Group distributionally robust optimization (group DRO) addresses this problem by minimizing the worst-group loss, but it fails when group losses misrepresent performance differences between groups. This is common in domains like speech, where the widely used connectionist temporal classification (CTC) loss not only scales with input length but also varies with linguistic and acoustic properties, leading to spurious differences between group losses. We present CTC-DRO, which addresses the shortcomings of the group DRO objective by smoothing the group weight update to prevent overemphasis on consistently high-loss groups, while using input length-matched batching to mitigate CTC's scaling issues. We evaluate CTC-DRO on the task of multilingual automatic speech recognition (ASR) across five language sets from the diverse ML-SUPERB 2.0 benchmark. CTC-DRO consistently outperforms group DRO and CTC-based baseline models, reducing the worst-language error by up to 47.1% and the average error by up to 32.9%. CTC-DRO can be applied to ASR with minimal computational costs, and, while motivated by multilingual ASR, offers the potential for reducing group disparities in other domains with similar challenges.
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。