A method of supervised learning from conflicting data with hidden contexts

Tianren Zhang, Yizhou Jiang, Feng Chen

Conventional supervised learning assumes a stable input-output relationship. However, this assumption fails in open-ended training settings where the input-output relationship depe…