FastInject：将未配对文本数据注入基于CTC的ASR训练
FastInject: Injecting Unpaired Text Data into CTC-based ASR training
AI总结：
针对基于CTC的端到端ASR模型的域偏移问题，提出FastInject平启动联合训练方法，通过预上采样文本单元和AM3机制注入多域未配对文本数据，有效提升了域内和域外识别性能。
摘要：
近年来，基于连接主义时间分类（CTC）的端到端（E2E）自动语音识别（ASR）模型取得了令人瞩目的成果，尤其是随着自监督学习的发展。然而，在配对语音-文本数据上训练的E2E ASR模型通常会遭遇从训练到测试的域偏移问题。为缓解这一问题，本文提出一种名为FastInject的平启动联合训练方法，可高效地将多域未配对文本数据注入基于CTC的ASR训练。为保持训练效率，文本单元被预上采样，其表征与语音特征一同被输入CTC模型。为弥合语音与文本之间的模态鸿沟，本文提出了基于注意力的模态匹配机制（AM3），该机制保留了E2E平启动训练特性。实验表明，所提出的FastInject在域内Librispeech-100h数据集上实现了22%的相对词错误率降低（WERR），在域外测试集上实现了20%的相对WERR。
英文摘要：
Recently, connectionist temporal classification (CTC)-based end-to-end (E2E) automatic speech recognition (ASR) models have achieved impressive results, especially with the development of self-supervised learning. However, E2E ASR models trained on paired speech-text data often suffer from domain shifts from training to testing. To alleviate this issue, this paper proposes a flat-start joint training method, named FastInject, which efficiently injects multi-domain unpaired text data into CTC-based ASR training. To maintain training efficiency, text units are pre-upsampled, and their representations are fed into the CTC model along with speech features. To bridge the modality gap between speech and text, an attention-based modality matching mechanism (AM3) is proposed, which retains the E2E flat-start training. Experiments show that the proposed FastInject gave a 22\% relative WER reduction (WERR) for intra-domain Librispeech-100h data and 20\% relative WERR on out-of-domain test sets.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
