一篇新发表在arXiv上的研究论文介绍了一种用于自动语音识别（ASR）系统的双参考基准测试方法，特别解决了非典型语音的挑战。研究强调，大多数ASR评估将逐字转录和意图转录参考混淆，可能错误地表示模型性能。通过在口吃语音上使用逐字和意图参考对11个ASR模型进行基准测试，研究表明模型排名根据所选参考风格的不同存在显著差异。这突显了为准确的模型评估选择适当的转录参考的至关重要性，尤其是在涉及非典型语音的用例中。 AI
影响 强调了对ASR系统进行细致评估的必要性，可能影响未来非典型语音的开发和基准测试标准。
排序理由 发表在arXiv上的研究论文，介绍了一种用于ASR系统的新基准测试方法。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2606.31112v1 Announce Type: new Abstract: ASR systems have been often reported to underperform on atypical speech. An often conflated compounding factor is the existence of two valid transcription references: verbatim (actual produced speech, including repetitions/prolongat…
ASR systems have been often reported to underperform on atypical speech. An often conflated compounding factor is the existence of two valid transcription references: verbatim (actual produced speech, including repetitions/prolongations) and intended (the canonical form of the te…