Mitigating Bias Using Model-Agnostic Data Attribution

arXiv:2405.05031v1 Announce Type: new
Abstract: Mitigating bias in machine learning models is a critical endeavor for ensuring fairness and equity. In this paper, we propose a novel approach to address bias by leveraging pixel image attributions to identify and regularize regions of images containing significant information about bias attributes. Our method utilizes a model-agnostic approach to extract pixel attributions by employing a convolutional neural network (CNN) classifier trained on small image patches. By training the classifier to predict a property …

abstract arxiv attribution bias cs.cv data endeavor equity fairness identify image images information machine machine learning machine learning models model-agnostic novel paper pixel type