# Improving Speech Recognition on Noisy Speech via Speech Enhancement with Multi-Discriminators CycleGAN

arXiv: https://arxiv.org/abs/2112.06309 · HTML: https://mlsanity.com/paper/2112.06309/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 25 references that are in the map.
- First posted on arXiv 12 Dec 2021.

## Builds on (most cited first, 9 of 9)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Domain-Adversarial Training of Neural Networks](https://mlsanity.com/paper/1505.07818/) (2015)
- [Generative Multi-Adversarial Networks](https://mlsanity.com/paper/1611.01673/) (2016)
- [Noise Adaptive Speech Enhancement using Domain Adversarial Training](https://mlsanity.com/paper/1807.07501/) (2018)
- [A Multi-Discriminator CycleGAN for Unsupervised Non-Parallel Speech Domain Adaptation](https://mlsanity.com/paper/1804.00522/) (2018)
- [Improving End-to-End Speech Recognition with Policy Learning](https://mlsanity.com/paper/1712.07101/) (2017)
- [Adversarial Feature-Mapping for Speech Enhancement](https://mlsanity.com/paper/1809.02251/) (2018)
- [Cycle-Consistent Speech Enhancement](https://mlsanity.com/paper/1809.02253/) (2018)

## Abstract (arXiv)

This paper presents our latest investigations on improving automatic speech recognition for noisy speech via speech enhancement. We propose a novel method named Multi-discriminators CycleGAN to reduce noise of input speech and therefore improve the automatic speech recognition performance. Our proposed method leverages the CycleGAN framework for speech enhancement without any parallel data and improve it by introducing multiple discriminators that check different frequency areas. Furthermore, we show that training multiple generators on homogeneous subset of the training data is better than training one generator on all the training data. We evaluate our method on CHiME-3 data set and observe up to 10.03% relatively WER improvement on the development set and up to 14.09% on the evaluation set.