Reliability and Learnability of Human Bandit Feedback for Sequence-to-Sequence Reinforcement Learning

Julia Kreutzer1 and Joshua Uyheng3  and Stefan Riezler1;2

1Computational Linguistics & 2IWR, Heidelberg University, Germany

更多

fkreutzer,riezlerg@cl.uni-heidelberg.de

3Departments of Psychology & Mathematics, Ateneo de Manila University, Philippines

juyheng@ateneo.edu

收起