搜索
Home
»
machine-learning
» Addressing Delayed Feedback for Continuous Training with Neural Networks in CTR prediction
Addressing Delayed Feedback for Continuous Training with Neural Networks in CTR prediction
Table of Contents
摘要
摘要
解决的问题: CTR预估流式训练的时候,正样本反馈往往存在很大延时
原来的解决方案: 将所有样本都看做负样本,直到看到正样本的反馈