Confident Learning: Estimating Uncertainty in Dataset Labels

作者： Isaac L. Chuang , Lu Jiang , Curtis G. Northcutt

关键词: Pruning (decision trees) 、 Context (language use) 、 Focus (optics) 、 Probabilistic logic 、 Pattern recognition 、 Joint probability distribution 、 Ranking 、 Class (biology) 、 Noise (video) 、 Artificial intelligence 、 Computer science

摘要: … Second, we seek to understand how CL may improve learning with noise in the label space of text data, but not noise in the text data itself (eg typos). Towards this goal, we consider non…