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.

该论文提出了一种新颖的非自回归（NAR）基于块的注意力掩码解码器（AMD），用于平衡 Conformer ASR 系统的性能效率权衡，实现在输出标签的连续块内并行 NAR 推理，通过使用注意力掩码隐藏，同时在块之间进行左到右的 AR 预测和历史上下文融合。设计了一种波束搜索算法，利用 CTC、AR 解码器和 AMD 概率的动态融合。LibriSpeech-100hr 语料库上的实验证明，结合 AMD 模块的三部分解码器在基线 CTC+AR 解码上实现了最大 1.73 倍的解码加速比，而在测试集上未引起统计学显著的词错误率（WER）增加。在具有相同解码实时因素的情况下，相对于 CTC+AR 基线，实现了高达 0.7% 和 0.3% 绝对（相对 5.3% 和 6.1%）的统计学显著 WER 降低。