# Bias-Aware Loss for Training Image and Speech Quality Prediction Models from Multiple Datasets

arXiv: https://arxiv.org/abs/2104.10217 · HTML: https://mlsanity.com/paper/2104.10217/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 23 references that are in the map.
- First posted on arXiv 20 Apr 2021.

## Builds on (most cited first, 6 of 6)

- [Deep Neural Networks for No-Reference and Full-Reference Image Quality Assessment](https://mlsanity.com/paper/1612.01697/) (2016)
- [Massive Online Crowdsourced Study of Subjective and Objective Picture Quality](https://mlsanity.com/paper/1511.02919/) (2015)
- [NISQA: A Deep CNN-Self-Attention Model for Multidimensional Speech Quality Prediction with Crowdsourced Datasets](https://mlsanity.com/paper/2104.09494/) (2021)
- [MOSNet: Deep Learning based Objective Assessment for Voice Conversion](https://mlsanity.com/paper/1904.08352/) (2019)
- [RAN4IQA: Restorative Adversarial Nets for No-Reference Image Quality Assessment](https://mlsanity.com/paper/1712.05444/) (2017)
- [Transformation of Mean Opinion Scores to Avoid Misleading of Ranked based Statistical Techniques](https://mlsanity.com/paper/2004.11490/) (2020)

## Built on by (most cited first, 3 of 3)

- [LAR-IQA: A Lightweight, Accurate, and Robust No-Reference Image Quality Assessment Model](https://mlsanity.com/paper/2408.17057/) (2024)
- [From Scores to Preferences: Redefining MOS Benchmarking for Speech Quality Reward Modeling](https://mlsanity.com/paper/2510.00743/) (2025)
- [Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities](https://mlsanity.com/paper/2206.13272/) (2022)

## Abstract (arXiv)

The ground truth used for training image, video, or speech quality prediction models is based on the Mean Opinion Scores (MOS) obtained from subjective experiments. Usually, it is necessary to conduct multiple experiments, mostly with different test participants, to obtain enough data to train quality models based on machine learning. Each of these experiments is subject to an experiment-specific bias, where the rating of the same file may be substantially different in two experiments (e.g. depending on the overall quality distribution). These different ratings for the same distortion levels confuse neural networks during training and lead to lower performance. To overcome this problem, we propose a bias-aware loss function that estimates each dataset's biases during training with a linear function and considers it while optimising the network weights. We prove the efficiency of the proposed method by training and validating quality prediction models on synthetic and subjective image and speech quality datasets.