https://ai-scholar.tech/articles/reinforcement-learning/GTrXL
ʂ̃TCgɃWv悤ƂĂ܂BXΏL̃NNbNĂ
UPLIFTōLȂő̌܂傤IKȉ{Ct܂I