用熵最小化和说话人编码，仅用1分钟语音就能稳定适配语音识别器。
Robust Unsupervised Adaptation of a Speech Recogniser Using Entropy Minimisation and Speaker Codes
- 用多假设条件熵替代单伪标签交叉熵，提升适应鲁棒性。
- 在Common Voice的远场噪声数据上，1分钟适配使误字率降20%。
- 适合数据稀缺场景下的语音识别模型快速自适应。
语音识别器通常只在特定环境下表现良好，需适应新环境。针对新说话人适应时数据不足且无标签的问题，本文提出一种结合熵最小化与说话人编码的方法，使仅用1分钟数据的适应仍具鲁棒性。首先，采用基于完整假设的条件熵损失函数，避免依赖单一错误的识别假设；其次，引入短向量形式的“说话人编码”，可低数据量估计。在远场噪声增强版Common Voice数据集上，该方法在1分钟适配数据下实现相对20%的误字率降低，10分钟时提升至29%。
原文摘要 · Abstract (English)
Speech recognisers usually perform optimally only in a specific environment and need to be adapted to work well in another. For adaptation to a new speaker, there is often too little data for fine-tuning to be robust, and that data is usually unlabelled. This paper proposes a combination of approaches to make adaptation to a single minute of data robust. First, instead of estimating the adaptation parameters with cross-entropy on a single error-prone hypothesis or "pseudo-label", this paper proposes a novel loss function, the conditional entropy over complete hypotheses. Using multiple hypotheses makes adaptation more robust to errors in the initial recognition. Second, a "speaker code" characterises a speaker in a vector short enough that it requires little data to estimate. On a far-field noise-augmented version of Common Voice, the proposed scheme yields a 20% relative improvement in word error rate on one minute of adaptation data, increasing on 10 minutes to 29%.
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。