Dynamic Reward Adjustment in Multi-Reward Reinforcement Learning for Counselor Reflection Generation

arXiv:2403.13578v1 Announce Type: cross
Abstract: In this paper, we study the problem of multi-reward reinforcement learning to jointly optimize for multiple text qualities for natural language generation. We focus on the task of counselor reflection generation, where we optimize the generators to simultaneously improve the fluency, coherence, and reflection quality of generated counselor responses. We introduce two novel bandit methods, DynaOpt and C-DynaOpt, which rely on the broad strategy of combining rewards into a single value and optimizing them simultaneously. …

abstract arxiv cs.cl cs.lg dynamic focus generators language language generation multiple natural natural language natural language generation paper reinforcement reinforcement learning study text type