# Using RLHF to align speech enhancement approaches to mean-opinion quality scores

arXiv: https://arxiv.org/abs/2410.13182 · HTML: https://mlsanity.com/paper/2410.13182/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Oct 2024.

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

- [Multi-Metric Preference Alignment for Generative Speech Restoration](https://mlsanity.com/paper/2508.17229/) (2025)
- [Aligning Generative Speech Enhancement with Perceptual Feedback](https://mlsanity.com/paper/2507.09929/) (2025)
- [JSQA: Speech Quality Assessment with Perceptually-Inspired Contrastive Pretraining Based on JND Audio Pairs](https://mlsanity.com/paper/2507.11636/) (2025)
- [LLM-Guided Reinforcement Learning for Audio-Visual Speech Enhancement](https://mlsanity.com/paper/2603.13952/) (2026)
- [DAMOS: Learning Distortion-Aware Speech Quality Assessment through Explicit Distortion Localization](https://mlsanity.com/paper/2608.21176/) (2026)

## Abstract (arXiv)

Objective speech quality measures are typically used to assess speech enhancement algorithms, but it has been shown that they are sub-optimal as learning objectives because they do not always align well with human subjective ratings. This misalignment often results in noticeable distortions and artifacts that cause speech enhancement to be ineffective. To address these issues, we propose a reinforcement learning from human feedback (RLHF) framework to fine-tune an existing speech enhancement approach by optimizing performance using a mean-opinion score (MOS)-based reward model. Our results show that the RLHF-finetuned model has the best performance across different benchmarks for both objective and MOS-based speech quality assessment metrics on the Voicebank+DEMAND dataset. Through ablation studies, we show that both policy gradient loss and supervised MSE loss are important for balanced optimization across the different metrics.