Fairness in Deep Learning: A Computational Perspective

Fairness in deep learning has attracted tremendous attention recently, as deep learning is increasingly being used in high-stake decision making applications that affect individual lives. We provide a review covering recent progresses to tackle algorithmic fairness problems of deep learning from the computational perspective. Specifically, we show that interpretability can serve as a useful ingredient to diagnose the reasons that lead to algorithmic discrimination. We also discuss fairness mitigation approaches categorized according to three stages of deep learning life-cycle, aiming to push forward the area of fairness in deep learning and build genuinely fair and reliable deep learning systems.

Authors

• Mengnan Du New Jersey Institute of Technology
• Fan Yang Nanyang Technological University
• Na Zou University of Houston
• Xia Hu Rice University

Keywords

• Predictive models
• Deep learning
• Modeling
• Measurement
• Face recognition
• Neurons
• Computational Perspective
• Learning Algorithms
• Deep Neural Network
• Deep Learning Models
• Criminal Justice
• Protective Properties
• Prediction Task
• Dark Skin
• Intermediate Representation
• Algorithmic Bias
• Prediction Model
• Training Data
• Transfer Learning
• Generative Adversarial Networks
• Final Prediction
• Prediction Quality
• Demographic Groups
• Feature Subset
• Deep Neural Network Model
• Local Translation
• Sensitive Attributes
• Global Interpretation
• Class Activation Maps
• White Skin
• Adversarial Training
• Numerical Score
• Deep Representation
• Biased Representation
• DNN
• Fairness
• Bias
• Interpretability