Document worth reading: “Fairness in Deep Learning: A Computational Perspective”

Deep learning is increasingly getting used in high-stake willpower making capabilities that impact specific individual lives. However, deep learning fashions may exhibit algorithmic discrimination behaviors with respect to protected groups, doubtlessly posing damaging impacts on individuals and society. Therefore, fairness in deep learning has attracted giant consideration simply these days. We current an entire analysis overlaying current methods to kind out algorithmic fairness points from the computational perspective. Specifically, we current that interpretability can perform a useful ingredient, which is perhaps augmented into the biases detection and mitigation pipelines. We moreover deal with open evaluation points and future evaluation directions, aiming to push forward the realm of fairness in deep learning and assemble genuinely truthful, accountable, and clear deep learning methods. Fairness in Deep Learning: A Computational Perspective