Can Fairness be Automated? Guidelines and Opportunities for Fairness-aware AutoML

Abstract

The field of automated machine learning (AutoML) introduces techniques that automate parts of the development of machine learning (ML) systems, accelerating the process and reducing barriers for novices. However, decisions derived from ML models can reproduce, amplify, or even introduce unfairness in our societies, causing harm to (groups of) individuals. In response, researchers have started to propose AutoML systems that jointly optimize fairness and predictive performance to mitigate fairness-related harm. However, fairness is a complex and inherently interdisciplinary subject, and solely posing it as an optimization problem can have adverse side effects. With this work, we aim to raise awareness among developers of AutoML systems about such limitations of fairness-aware AutoML, while also calling attention to the potential of AutoML as a tool for fairness research. We present a comprehensive overview of different ways in which fairness-related harm can arise and the ensuing implications for the design of fairness-aware AutoML. We conclude that while fairness cannot be automated, fairness-aware AutoML can play an important role in the toolbox of ML practitioners. We highlight several open technical challenges for future work in this direction. Additionally, we advocate for the creation of more user-centered assistive systems designed to tackle challenges encountered in fairness work

References in corpus (37)

• Practical Bayesian Optimization of Machine Learning Algorithms
• Equality of Opportunity in Supervised Learning
• Fairness in Machine Learning: A Survey
• Scalable Bayesian Optimization Using Deep Neural Networks
• Do ImageNet Classifiers Generalize to ImageNet?
• Towards a Critical Race Methodology in Algorithmic Fairness
• Non-stochastic Best Arm Identification and Hyperparameter Optimization
• Runaway Feedback Loops in Predictive Policing
• Importance of Tuning Hyperparameters of Machine Learning Algorithms
• Multi-Objective Hyperparameter Optimization in Machine Learning -- An Overview
• Exploring How Machine Learning Practitioners (Try To) Use Fairness Toolkits
• Auto-WEKA: Combined Selection and Hyperparameter Optimization of Classification Algorithms
• Bayesian Optimization in AlphaGo
• Neural Architecture Search: Insights from 1000 Papers
• Sensible AI: Re-imagining Interpretability and Explainability using Sensemaking Theory
• A Comprehensive Survey on Hardware-Aware Neural Architecture Search
• Fairlearn: Assessing and Improving Fairness of AI Systems
• On the Moral Justification of Statistical Parity
• Automating Data Science: Prospects and Challenges
• Accounting for Variance in Machine Learning Benchmarks
• It's COMPASlicated: The Messy Relationship between RAI Datasets and Algorithmic Fairness Benchmarks
• On the Importance of Hyperparameter Optimization for Model-based Reinforcement Learning
• De-biasing "bias" measurement
• Operationalizing Machine Learning: An Interview Study
• AMLB: an AutoML Benchmark
• Promoting Fairness through Hyperparameter Optimization
• The four-fifths rule is not disparate impact: a woeful tale of epistemic trespassing in algorithmic fairness
• Assessing the Fairness of AI Systems: AI Practitioners' Processes, Challenges, and Needs for Support
• Rethinking Bias Mitigation: Fairer Architectures Make for Fairer Face Recognition
• Explaining Hyperparameter Optimization via Partial Dependence Plots
• Auto-PyTorch Tabular: Multi-Fidelity MetaLearning for Efficient and Robust AutoDL
• Multi-objective Asynchronous Successive Halving
• Mind the Gap: Measuring Generalization Performance Across Multiple Objectives
• Are There Exceptions to Goodhart's Law? On the Moral Justification of Fairness-Aware Machine Learning
• NAS-HPO-Bench-II: A Benchmark Dataset on Joint Optimization of Convolutional Neural Network Architecture and Training Hyperparameters
• FairAutoML: Embracing Unfairness Mitigation in AutoML
• Whither AutoML? Understanding the Role of Automation in Machine Learning Workflows