Knowledge Distillation from Multiple Foundation Models for End-to-End Speech Recognition
원문 초록
Although large foundation models pre-trained by self-supervised learning have achieved state-of-the-art performance in many tasks including automatic speech recognition (ASR), knowledge distillation (KD) is often required in practice to transfer the knowledge learned by large teacher models into much smaller student models with affordable computation and memory costs. This paper proposes a novel two-stage KD framework to distil the knowledge from multiple speech foundation models as teachers into a single student neural transducer model for ASR. In the first stage, the student model encoder is pre-trained using the embeddings extracted from multiple teacher models. In the second stage, the student encoder is fine-tuned with the audio-text pairs based on the ASR task. Experiments on the LibriSpeech 100-hour subset show that the proposed KD framework improves the performance of both streaming and non-streaming student models when using only one teacher. The performance of the student model can be further enhanced when multiple teachers are used jointly, achieving word error rate reductions (WERRs) of 17.5% and 10.6%. Our proposed framework can be combined with other existing KD methods to achieve further improvements. Further WERRs were obtained by incorporating extra unlabelled data during encoder pre-training, leading to a total relative WERR of 55.0% on the non-streaming student model.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Transformer Based Punctuation Restoration for Turkish
Explainability of Speech Recognition Transformers via Gradient-based Attention Visualization
Evaluating Variants of wav2vec 2.0 on Affective Vocal Burst Tasks
Mlphon: A Multifunctional Grapheme-Phoneme Conversion Tool Using Finite State Transducers
Building a Public Domain Voice Database for Odia
Automatic Speech Recognition in German: A Detailed Error Analysis
토론
이 논문과 연결된 포럼 글이 아직 없습니다.