# Generalizing RNN-Transducer to Out-Domain Audio via Sparse Self-Attention Layers

arXiv: https://arxiv.org/abs/2108.10752 · HTML: https://mlsanity.com/paper/2108.10752/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 35 references that are in the map.
- First posted on arXiv 22 Aug 2021.

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

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [Speech Recognition with Deep Recurrent Neural Networks](https://mlsanity.com/paper/1303.5778/) (2013)
- [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)
- [Deep Speech 2: End-to-End Speech Recognition in English and Mandarin](https://mlsanity.com/paper/1512.02595/) (2015)
- [Attention-Based Models for Speech Recognition](https://mlsanity.com/paper/1506.07503/) (2015)
- [State-of-the-art Speech Recognition With Sequence-to-Sequence Models](https://mlsanity.com/paper/1712.01769/) (2017)
- [Recent Developments on ESPnet Toolkit Boosted by Conformer](https://mlsanity.com/paper/2010.13956/) (2020)
- [A Streaming On-Device End-to-End Model Surpassing Server-Side Conventional Model Quality and Latency](https://mlsanity.com/paper/2003.12710/) (2020)
- [Exploring Architectures, Data and Units For Streaming End-to-End Speech Recognition with RNN-Transducer](https://mlsanity.com/paper/1801.00841/) (2018)
- [Monotonic Chunkwise Attention](https://mlsanity.com/paper/1712.05382/) (2017)
- [A Better and Faster End-to-End Model for Streaming ASR](https://mlsanity.com/paper/2011.10798/) (2020)
- [RWTH ASR Systems for LibriSpeech: Hybrid vs Attention -- w/o Data Augmentation](https://mlsanity.com/paper/1905.03072/) (2019)
- [Explicit Sparse Transformer: Concentrated Attention Through Explicit Selection](https://mlsanity.com/paper/1912.11637/) (2019)
- [Neural Speech Recognizer: Acoustic-to-Word LSTM Model for Large Vocabulary Speech Recognition](https://mlsanity.com/paper/1610.09975/) (2016)
- [RNN-T Models Fail to Generalize to Out-of-Domain Audio: Causes and Solutions](https://mlsanity.com/paper/2005.03271/) (2020)
- [Accelerating RNN Transducer Inference via One-Step Constrained Beam Search](https://mlsanity.com/paper/2002.03577/) (2020)
- [Building competitive direct acoustics-to-word models for English conversational speech recognition](https://mlsanity.com/paper/1712.03133/) (2017)
- [Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition](https://mlsanity.com/paper/2106.09236/) (2021)
- [Capturing Multi-Resolution Context by Dilated Self-Attention](https://mlsanity.com/paper/2104.02858/) (2021)

## Built on by (most cited first, 2 of 2)

- [Pruning Self-Attention for Zero-Shot Multi-Speaker Text-to-Speech](https://mlsanity.com/paper/2308.14909/) (2023)
- [MambAttention: Mamba with Multi-Head Attention for Generalizable Single-Channel Speech Enhancement](https://mlsanity.com/paper/2507.00966/) (2025)

## Abstract (arXiv)

Recurrent neural network transducer (RNN-T) is an end-to-end speech recognition framework converting input acoustic frames into a character sequence. The state-of-the-art encoder network for RNN-T is the Conformer, which can effectively model the local-global context information via its convolution and self-attention layers. Although Conformer RNN-T has shown outstanding performance, most studies have been verified in the setting where the train and test data are drawn from the same domain. The domain mismatch problem for Conformer RNN-T has not been intensively investigated yet, which is an important issue for the product-level speech recognition system. In this study, we identified that fully connected self-attention layers in the Conformer caused high deletion errors, specifically in the long-form out-domain utterances. To address this problem, we introduce sparse self-attention layers for Conformer-based encoder networks, which can exploit local and generalized global information by pruning most of the in-domain fitted global connections. Also, we propose a state reset method for the generalization of the prediction network to cope with long-form utterances. Applying proposed methods to an out-domain test, we obtained 27.6% relative character error rate (CER) reduction compared to the fully connected self-attention layer-based Conformers.