Towards Effective and Efficient Non-autoregressive Decoding Using Block-based Attention Mask
This work addresses efficiency improvements in automatic speech recognition for real-time applications, representing an incremental advance in non-autoregressive methods.
The paper tackles the performance-efficiency trade-off in non-autoregressive decoding for Conformer ASR systems by proposing a block-based Attention Mask Decoder, achieving a maximum decoding speed-up of 1.73x with no significant WER increase and up to 0.7% absolute WER reduction at the same speed.
This paper proposes a novel non-autoregressive (NAR) block-based Attention Mask Decoder (AMD) that flexibly balances performance-efficiency trade-offs for Conformer ASR systems. AMD performs parallel NAR inference within contiguous blocks of output labels that are concealed using attention masks, while conducting left-to-right AR prediction and history context amalgamation between blocks. A beam search algorithm is designed to leverage a dynamic fusion of CTC, AR Decoder, and AMD probabilities. Experiments on the LibriSpeech-100hr corpus suggest the tripartite Decoder incorporating the AMD module produces a maximum decoding speed-up ratio of 1.73x over the baseline CTC+AR decoding, while incurring no statistically significant word error rate (WER) increase on the test sets. When operating with the same decoding real time factors, statistically significant WER reductions of up to 0.7% and 0.3% absolute (5.3% and 6.1% relative) were obtained over the CTC+AR baseline.