# Multi-Level Modeling Units for End-to-End Mandarin Speech Recognition

arXiv: https://arxiv.org/abs/2205.11998 · HTML: https://mlsanity.com/paper/2205.11998/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 35 references that are in the map.
- First posted on arXiv 24 May 2022.

## Builds on (most cited first, 24 of 24)

- [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)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [Listen, Attend and Spell](https://mlsanity.com/paper/1508.01211/) (2015)
- [Attention-Based Models for Speech Recognition](https://mlsanity.com/paper/1506.07503/) (2015)
- [AISHELL-1: An Open-Source Mandarin Speech Corpus and A Speech Recognition Baseline](https://mlsanity.com/paper/1709.05522/) (2017)
- [Recent Developments on ESPnet Toolkit Boosted by Conformer](https://mlsanity.com/paper/2010.13956/) (2020)
- [WeNet: Production oriented Streaming and Non-streaming End-to-End Speech Recognition Toolkit](https://mlsanity.com/paper/2102.01547/) (2021)
- [Intermediate Loss Regularization for CTC-based Speech Recognition](https://mlsanity.com/paper/2102.03216/) (2021)
- [Relaxing the Conditional Independence Assumption of CTC-based ASR by Conditioning on Intermediate Predictions](https://mlsanity.com/paper/2104.02724/) (2021)
- [Transformers with convolutional context for ASR](https://mlsanity.com/paper/1904.11660/) (2019)
- [Multitask Learning with Low-Level Auxiliary Tasks for Encoder-Decoder Based Speech Recognition](https://mlsanity.com/paper/1704.01631/) (2017)
- [Improving RNN Transducer Based ASR with Auxiliary Tasks](https://mlsanity.com/paper/2011.03109/) (2020)
- [Deja-vu: Double Feature Presentation and Iterated Loss in Deep Transformer Networks](https://mlsanity.com/paper/1910.10324/) (2019)
- [Syllable-Based Sequence-to-Sequence Speech Recognition with the Transformer in Mandarin Chinese](https://mlsanity.com/paper/1804.10752/) (2018)
- [Hierarchical Multi Task Learning With CTC](https://mlsanity.com/paper/1807.07104/) (2018)
- [Cascade RNN-Transducer: Syllable Based Streaming On-device Mandarin Speech Recognition with a Syllable-to-Character Converter](https://mlsanity.com/paper/2011.08469/) (2020)
- [Hierarchical Multitask Learning for CTC-based Speech Recognition](https://mlsanity.com/paper/1807.06234/) (2018)
- [A Comparison of Modeling Units in Sequence-to-Sequence Speech Recognition with the Transformer on Mandarin Chinese](https://mlsanity.com/paper/1805.06239/) (2018)
- [Non-autoregressive Mandarin-English Code-switching Speech Recognition](https://mlsanity.com/paper/2104.02258/) (2021)
- [Consistent Training and Decoding For End-to-end Speech Recognition Using Lattice-free MMI](https://mlsanity.com/paper/2112.02498/) (2021)
- [Decoupling recognition and transcription in Mandarin ASR](https://mlsanity.com/paper/2108.01129/) (2021)
- [A comparable study of modeling units for end-to-end Mandarin speech recognition](https://mlsanity.com/paper/1805.03832/) (2018)

## Abstract (arXiv)

The choice of modeling units is crucial for automatic speech recognition (ASR) tasks. In mandarin scenarios, the Chinese characters represent meaning but are not directly related to the pronunciation. Thus only considering the writing of Chinese characters as modeling units is insufficient to capture speech features. In this paper, we present a novel method involves with multi-level modeling units, which integrates multi-level information for mandarin speech recognition. Specifically, the encoder block considers syllables as modeling units and the decoder block deals with character-level modeling units. To facilitate the incremental conversion from syllable features to character features, we design an auxiliary task that applies cross-entropy (CE) loss to intermediate decoder layers. During inference, the input feature sequences are converted into syllable sequences by the encoder block and then converted into Chinese characters by the decoder block. Experiments on the widely used AISHELL-1 corpus demonstrate that our method achieves promising results with CER of 4.1%/4.6% and 4.6%/5.2%, using the Conformer and the Transformer backbones respectively.