# Understanding Bias in Machine Learning

**Authors:** Jindong Gu, Daniela Oelke  
**arXiv:** [1909.01866](https://arxiv.org/abs/1909.01866)  
**Date:** 2019-09-02  
**Summary page:** https://www.summarizepaper.com/en/paper/1909.01866v1/

## Key points

- Bias is a problem in many domains such as human resources, public sector, and healthcare
- Machine learning was hoped to diminish or resolve this problem
- Experts warn that machine learning models can also be biased
- Jindong Gu and Daniela Oelke wrote an article "Understanding Bias in Machine Learning" explaining the issue of bias in machine learning from a technical perspective
- Biased data can have an impact on a machine learning model
- The authors developed interactive plots to visualize the bias learned from synthetic data
- Understanding and addressing bias in machine learning is crucial for ensuring fairness and avoiding unintended consequences
- Ethical considerations are important when developing and deploying machine learning models to ensure fairness and prevent negative outcomes.

## Summary

In many domains such as human resources, the public sector, and healthcare, bias is known to impede fair decisions. While there has been hope that machine learning methods could diminish or even resolve this problem, experts warn that machine learning models can also be biased. In their article "Understanding Bias in Machine Learning," Jindong Gu and Daniela Oelke explain the issue of bias in machine learning from a technical perspective and illustrate the impact that biased data can have on a machine learning model. To achieve this goal, they develop interactive plots to visualize the bias learned from synthetic data. The authors emphasize that understanding and addressing bias in machine learning is crucial for ensuring fairness and avoiding unintended consequences. Their work highlights the importance of ethical considerations when developing and deploying machine learning models to ensure fairness and prevent any negative outcomes.