# Non-autoregressive Transformer with Unified Bidirectional Decoder for Automatic Speech Recognition

arXiv: https://arxiv.org/abs/2109.06684 · HTML: https://mlsanity.com/paper/2109.06684/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 30 references that are in the map.
- First posted on arXiv 14 Sep 2021.

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [ESPnet: End-to-End Speech Processing Toolkit](https://mlsanity.com/paper/1804.00015/) (2018)
- [Non-Autoregressive Neural Machine Translation](https://mlsanity.com/paper/1711.02281/) (2017)
- [AISHELL-1: An Open-Source Mandarin Speech Corpus and A Speech Recognition Baseline](https://mlsanity.com/paper/1709.05522/) (2017)
- [Scaling Neural Machine Translation](https://mlsanity.com/paper/1806.00187/) (2018)
- [State-of-the-art Speech Recognition With Sequence-to-Sequence Models](https://mlsanity.com/paper/1712.01769/) (2017)
- [Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement](https://mlsanity.com/paper/1802.06901/) (2018)
- [Mask CTC: Non-Autoregressive End-to-End ASR with CTC and Mask Predict](https://mlsanity.com/paper/2005.08700/) (2020)
- [Aligned Cross Entropy for Non-Autoregressive Machine Translation](https://mlsanity.com/paper/2004.01655/) (2020)
- [Towards Fast and Accurate Streaming End-to-End ASR](https://mlsanity.com/paper/2004.11544/) (2020)
- [U2++: Unified Two-pass Bidirectional End-to-end Model for Speech Recognition](https://mlsanity.com/paper/2106.05642/) (2021)
- [Spike-Triggered Non-Autoregressive Transformer for End-to-End Speech Recognition](https://mlsanity.com/paper/2005.07903/) (2020)
- [CASS-NAT: CTC Alignment-based Single Step Non-autoregressive Transformer for Speech Recognition](https://mlsanity.com/paper/2010.14725/) (2020)
- [Non-Autoregressive Transformer ASR with CTC-Enhanced Decoder Input](https://mlsanity.com/paper/2010.15025/) (2020)
- [Listen Attentively, and Spell Once: Whole Sentence Generation via a Non-Autoregressive Architecture for Low-Latency Speech Recognition](https://mlsanity.com/paper/2005.04862/) (2020)
- [Insertion-Based Modeling for End-to-End Automatic Speech Recognition](https://mlsanity.com/paper/2005.13211/) (2020)
- [Transformer with Bidirectional Decoder for Speech Recognition](https://mlsanity.com/paper/2008.04481/) (2020)
- [TSNAT: Two-Step Non-Autoregressvie Transformer Models for Speech Recognition](https://mlsanity.com/paper/2104.01522/) (2021)

## Built on by (most cited first, 5 of 5)

- [Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognition](https://mlsanity.com/paper/2206.08317/) (2022)
- [A Survey on Non-Autoregressive Generation for Neural Machine Translation and Beyond](https://mlsanity.com/paper/2204.09269/) (2022)
- [Speech-text based multi-modal training with bidirectional attention for improved speech recognition](https://mlsanity.com/paper/2211.00325/) (2022)
- [Rethinking Speech Recognition with A Multimodal Perspective via Acoustic and Semantic Cooperative Decoding](https://mlsanity.com/paper/2305.14049/) (2023)
- [Audio-Conditioned Diffusion LLMs for ASR and Deliberation Processing](https://mlsanity.com/paper/2509.16622/) (2025)

## Abstract (arXiv)

Non-autoregressive (NAR) transformer models have been studied intensively in automatic speech recognition (ASR), and a substantial part of NAR transformer models is to use the casual mask to limit token dependencies. However, the casual mask is designed for the left-to-right decoding process of the non-parallel autoregressive (AR) transformer, which is inappropriate for the parallel NAR transformer since it ignores the right-to-left contexts. Some models are proposed to utilize right-to-left contexts with an extra decoder, but these methods increase the model complexity. To tackle the above problems, we propose a new non-autoregressive transformer with a unified bidirectional decoder (NAT-UBD), which can simultaneously utilize left-to-right and right-to-left contexts. However, direct use of bidirectional contexts will cause information leakage, which means the decoder output can be affected by the character information from the input of the same position. To avoid information leakage, we propose a novel attention mask and modify vanilla queries, keys, and values matrices for NAT-UBD. Experimental results verify that NAT-UBD can achieve character error rates (CERs) of 5.0%/5.5% on the Aishell1 dev/test sets, outperforming all previous NAR transformer models. Moreover, NAT-UBD can run 49.8x faster than the AR transformer baseline when decoding in a single step.