FairLay-ML: Intuitive Remedies for Unfairness in Data-Driven Social-Critical Algorithms
Open Access
- Author:
- Yu, Normen
- Area of Honors:
- Computer Science
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- G. Gary Tan, Thesis Supervisor
Danfeng Zhang, Thesis Honors Advisor - Keywords:
- machine learning
fairness in AI
human machine interface - Abstract:
- This thesis explores open-sourced machine learning (ML) model explanation tools to understand whether these tools can allow a layman to visualize, understand, and suggest intuitive remedies to unfairness in ML-based decision-support systems. Machine learning models trained on datasets biased against minority groups are increasingly used to guide life-altering social decisions, prompting the urgent need to study their logic for unfairness. Due to this problem's impact on vast populations of the general public, it is critical for the layperson -- not just subject matter experts in social justice or machine learning experts -- to understand the nature of unfairness within these algorithms and the potential trade-offs. Existing research on fairness in machine learning focuses mostly on the mathematical definitions and tools to understand and remedy unfair models, with some directly citing user-interactive tools as necessary for future work. This thesis presents FairLay-ML, a proof-of-concept GUI integrating some of the most promising tools to provide intuitive explanations for unfair logic in ML models by integrating existing research tools (e.g. Local Interpretable Model-Agnostic Explanations) with existing ML-focused GUI (e.g. Python Streamlit). We test FairLay-ML using models of various accuracy and fairness generated by an unfairness detector tool, Parfait-ML, and validate our results using Themis. Our study finds that the technology stack used for FairLay-ML makes it easy to install and provides real-time black-box explanations of pre-trained models to users. Furthermore, the explanations provided translate to actionable remedies. Out of the twenty-four unfair models studied, we are able to provide a very clear explanation to four. Of the four, three lead to a clear increase in fairness for age, gender, and race across the models without decrease in accuracy. For example, FairLay-ML indicates that native country is used as a proxy to determine someone's race in one of the unfair models. In this example, FairLay-ML indicates that someone from South Africa is very likely not to be Caucasian, and the model decreases its prediction probability by 0.02 for someone from South Africa. We show that masking native country leads to a fairer model.