论文阅读3-----基于强化学习的推荐系统 Top-K Off-Policy Correction for a REINFORCE Recommender System
abstract
problems in recommendation: a complex user state space (但好在有很多隐式的数据可以使用)
the problem of previous versions of the recommender: only observing feedback on recommendations selected by the previous versions of the recommender(只考虑正反馈,可考虑负反馈或是被忽视的item)
所以作者提出了以强化学习为基础的推荐系统(可以考虑负反馈以及其他的反馈)
contribuation
(1)scaling REINFORCE to a production receommender system with an action space on the orders of millions.(就是可以适应huge action space)
(2)applying off-policy correction to address data biases in learning from logged feedback collected form multiple behavior policies(利用importance sampling方法实现离线训练,看不懂转李宏毅力强化学习,看PPO哈)
(3)proposing a novel top-k off-policy correction to account f