Generalizing Orthogonalization for Models with Non-linearities

arXiv:2405.02475v1 Announce Type: new
Abstract: The complexity of black-box algorithms can lead to various challenges, including the introduction of biases. These biases present immediate risks in the algorithms' application. It was, for instance, shown that neural networks can deduce racial information solely from a patient's X-ray scan, a task beyond the capability of medical experts. If this fact is not known to the medical expert, automatic decision-making based on this algorithm could lead to prescribing a treatment (purely) based on …

abstract algorithms application arxiv beyond biases box capability challenges complexity cs.ai cs.lg information instance introduction medical networks neural networks patient racial ray risks stat.co stat.me type x-ray