arXiv cs.CL· Kien Nguyen, Tuan Dinh, Bao Phan·· 11 小时前AI 评分34
微调 Transducer 改善受损语音识别：SAPC2 挑战赛的增益分布与评测启示
Fine-Tuning a Transducer for Impaired Speech: Where the Gains Appear and What They Mean for Evaluation
AI 导读
在 Speech Accessibility Project Challenge 2（SAPC2）中，研究者用公开语音识别器在挑战数据上微调，最佳成绩在隐藏测试集 Test1 上取得 7.81% 字符错误率，官方基线为 15.26%。
正文 · 原文
Abstract:Speech recognition still works poorly for many people whose speech is affected by conditions such as ALS, Parkinson's disease, cerebral palsy, Down syndrome or stroke. We took part in the Speech Accessibility Project Challenge 2 (SAPC2) with a public speech recognizer that we fine-tuned on the challenge data. Our best entry scored 7.81% character error rate on Test1, the first of two hidden test sets (official baseline 15.26%). Our live entries ran on CPUs with a combined median latency of about 0.9 seconds at 7.96-8.07% error. On new speakers, fine-tuning reduced errors most for the speakers that an off-the-shelf recognizer finds hardest to understand. Sentences already seen in training had far lower error rates than new ones (2.58% vs 6.42%), a larger difference than being a familiar speaker made. We close with lessons for evaluating such systems.
| Comments: | 6 pages, 4 tables. Accepted at the SAPC2 Competition Workshop at NeurIPS 2026 |
| Subjects: | Computation and Language (cs.CL) |
| ACM classes: | I.2.7 |
| Cite as: | arXiv:2611.07587 [cs.CL] |
| (or arXiv:2611.07587v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.07587 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Kien Nguyen [view email]
[v1]
Thu, 5 Nov 2026 07:39:12 UTC (13 KB)
来源：arXiv cs.CL · arxiv.org