We develop a framework for repairing machine learning models by identifying detrimental training datapoints and erasing their memories.
確定! 回上一頁