Towards Effective And Efficient Non-autoregressive Decoding Using Block-based Attention Mask

Abstract

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.

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keywang2024towards

Related papers

• Improved Mask-ctc For Non-autoregressive End-to-end ASR (2020)11.76
• Non-autoregressive Transformer ASR With Ctc-enhanced Decoder Input (2020)10.97
• Non-autoregressive End-to-end Approaches For Joint Automatic Speech Recognition And Spoken Language Understanding (2023)5.84
• Segment-level Vectorized Beam Search Based On Partially Autoregressive Inference (2023)0.00
• 4D ASR: Joint Modeling Of CTC, Attention, Transducer, And Mask-predict Decoders (2022)7.50
• Effectiveasr: A Single-step Non-autoregressive Mandarin Speech Recognition Architecture With High Accuracy And Inference Speed (2024)3.58
• Non-autoregressive Transformer With Unified Bidirectional Decoder For Automatic Speech Recognition (2021)7.81
• Effective Decoder Masking For Transformer Based End-to-end Speech Recognition (2020)0.00