# Remember to correct the bias when using deep learning for regression!

arXiv: https://arxiv.org/abs/2203.16470 · HTML: https://mlsanity.com/paper/2203.16470/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 31 references that are in the map.
- First posted on arXiv 30 Mar 2022.

## Builds on (most cited first, 8 of 8)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [MnasNet: Platform-Aware Neural Architecture Search for Mobile](https://mlsanity.com/paper/1807.11626/) (2018)
- [Distribution Calibration for Regression](https://mlsanity.com/paper/1905.06023/) (2019)
- [Deep Learning Based 3D Point Cloud Regression for Estimating Forest Biomass](https://mlsanity.com/paper/2112.11335/) (2021)

## Built on by (most cited first, 4 of 4)

- [CHILI: Chemically-Informed Large-scale Inorganic Nanomaterials Dataset for Advancing Graph Machine Learning](https://mlsanity.com/paper/2402.13221/) (2024)
- [FlowMark: Mask-Guided Video Watermarking](https://mlsanity.com/paper/2607.05261/) (2026)
- [Transforming Hyperspectral Images Into Chemical Maps: A Novel End-to-End Deep Learning Approach](https://mlsanity.com/paper/2504.14131/) (2025)
- [Phase-OTDR Event Detection Using Image-Based Data Transformation and Deep Learning](https://mlsanity.com/paper/2512.05830/) (2025)

## Abstract (arXiv)

When training deep learning models for least-squares regression, we cannot expect that the training error residuals of the final model, selected after a fixed training time or based on performance on a hold-out data set, sum to zero. This can introduce a systematic error that accumulates if we are interested in the total aggregated performance over many data points. We suggest to adjust the bias of the machine learning model after training as a default postprocessing step, which efficiently solves the problem. The severeness of the error accumulation and the effectiveness of the bias correction is demonstrated in exemplary experiments.