Announcing The First Machine Learning Challenge

Deep learning has recently made remarkable progress in a wide range of applications, from image generation to language models. However, the widespread use of deep neural network models is tricky due to potential risks such as unfair bias propagation and amplification. Machine non-learning is an emerging subfield of machine learning that aims to remove the influence of a specific subset of training examples from a trained model, while preserving other beneficial properties. This article discusses the challenges of machine non-learning and the potential solutions to address them.