AILGMar 17, 2021

Transformer-based ASR Incorporating Time-reduction Layer and Fine-tuning with Self-Knowledge Distillation

arXiv:2103.09903v18 citations
Originality Incremental advance
AI Analysis

This work addresses computational efficiency and performance improvements for ASR systems, particularly in speech processing, though it appears incremental by building on existing Transformer-based methods.

The paper tackles the challenge of high computational complexity and poor semantic learning in Transformer-based automatic speech recognition (ASR) by proposing a time-reduction layer to lower frame-rates and a fine-tuning method with self-knowledge distillation, achieving state-of-the-art word error rate results on LibriSpeech datasets with 30 million parameters.

End-to-end automatic speech recognition (ASR), unlike conventional ASR, does not have modules to learn the semantic representation from speech encoder. Moreover, the higher frame-rate of speech representation prevents the model to learn the semantic representation properly. Therefore, the models that are constructed by the lower frame-rate of speech encoder lead to better performance. For Transformer-based ASR, the lower frame-rate is not only important for learning better semantic representation but also for reducing the computational complexity due to the self-attention mechanism which has O(n^2) order of complexity in both training and inference. In this paper, we propose a Transformer-based ASR model with the time reduction layer, in which we incorporate time reduction layer inside transformer encoder layers in addition to traditional sub-sampling methods to input features that further reduce the frame-rate. This can help in reducing the computational cost of the self-attention process for training and inference with performance improvement. Moreover, we introduce a fine-tuning approach for pre-trained ASR models using self-knowledge distillation (S-KD) which further improves the performance of our ASR model. Experiments on LibriSpeech datasets show that our proposed methods outperform all other Transformer-based ASR systems. Furthermore, with language model (LM) fusion, we achieve new state-of-the-art word error rate (WER) results for Transformer-based ASR models with just 30 million parameters trained without any external data.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes