# Joint Optimization of Streaming and Non-Streaming Automatic Speech Recognition with Multi-Decoder and Knowledge Distillation

arXiv: https://arxiv.org/abs/2405.13514 · HTML: https://mlsanity.com/paper/2405.13514/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Speech and audio distillation, one of 12 directions in Knowledge distillation.
- Builds on 23 of its 32 references that are in the map.
- First posted on arXiv 22 May 2024.

## Placed in

- Knowledge distillation / [Speech and audio distillation](https://mlsanity.com/distill/speech-and-audio-distillation/)

## Builds on (most cited first, 23 of 23)

- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [ESPnet: End-to-End Speech Processing Toolkit](https://mlsanity.com/paper/1804.00015/) (2018)
- [Listen, Attend and Spell](https://mlsanity.com/paper/1508.01211/) (2015)
- [Similarity-Preserving Knowledge Distillation](https://mlsanity.com/paper/1907.09682/) (2019)
- [Attention-Based Models for Speech Recognition](https://mlsanity.com/paper/1506.07503/) (2015)
- [Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learning](https://mlsanity.com/paper/1609.06773/) (2016)
- [Streaming End-to-end Speech Recognition For Mobile Devices](https://mlsanity.com/paper/1811.06621/) (2018)
- [Recent Developments on ESPnet Toolkit Boosted by Conformer](https://mlsanity.com/paper/2010.13956/) (2020)
- [End-to-End Speech Recognition: A Survey](https://mlsanity.com/paper/2303.03329/) (2023)
- [Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducer](https://mlsanity.com/paper/1801.00841/) (2018)
- [Branchformer: Parallel MLP-Attention Architectures to Capture Local and Global Context for Speech Recognition and Understanding](https://mlsanity.com/paper/2207.02971/) (2022)
- [Mask CTC: Non-Autoregressive End-to-End ASR with CTC and Mask Predict](https://mlsanity.com/paper/2005.08700/) (2020)
- [Streaming automatic speech recognition with the transformer model](https://mlsanity.com/paper/2001.02674/) (2020)
- [FastEmit: Low-latency Streaming ASR with Sequence-level Emission Regularization](https://mlsanity.com/paper/2010.11148/) (2020)
- [A Better and Faster End-to-End Model for Streaming ASR](https://mlsanity.com/paper/2011.10798/) (2020)
- [Cascaded encoders for unifying streaming and non-streaming ASR](https://mlsanity.com/paper/2010.14606/) (2020)
- [Towards Fast and Accurate Streaming End-to-End ASR](https://mlsanity.com/paper/2004.11544/) (2020)
- [Efficient Knowledge Distillation for RNN-Transducer Models](https://mlsanity.com/paper/2011.06110/) (2020)
- [Transformer ASR with Contextual Block Processing](https://mlsanity.com/paper/1910.07204/) (2019)
- [A Unified Cascaded Encoder ASR Model for Dynamic Model Sizes](https://mlsanity.com/paper/2204.06164/) (2022)
- [Conformer with dual-mode chunked attention for joint online and offline ASR](https://mlsanity.com/paper/2206.11157/) (2022)
- [Streaming parallel transducer beam search with fast-slow cascaded encoders](https://mlsanity.com/paper/2203.15773/) (2022)
- [4D ASR: Joint modeling of CTC, Attention, Transducer, and Mask-Predict decoders](https://mlsanity.com/paper/2212.10818/) (2022)

## Abstract (arXiv)

End-to-end (E2E) automatic speech recognition (ASR) can operate in two modes: streaming and non-streaming, each with its pros and cons. Streaming ASR processes the speech frames in real-time as it is being received, while non-streaming ASR waits for the entire speech utterance; thus, professionals may have to operate in either mode to satisfy their application. In this work, we present joint optimization of streaming and non-streaming ASR based on multi-decoder and knowledge distillation. Primarily, we study 1) the encoder integration of these ASR modules, followed by 2) separate decoders to make the switching mode flexible, and enhancing performance by 3) incorporating similarity-preserving knowledge distillation between the two modular encoders and decoders. Evaluation results show 2.6%-5.3% relative character error rate reductions (CERR) on CSJ for streaming ASR, and 8.3%-9.7% relative CERRs for non-streaming ASR within a single model compared to multiple standalone modules.