TY  - RPRT
TI  - Offline RL by Reward-Weighted Fine-Tuning for Conversation Optimization
AU  - Subhojyoti Mukherjee
AU  - Viet Dac Lai
AU  - Raghavendra Addanki
AU  - Ryan Rossi
AU  - Seunghyun Yoon
AU  - Trung Bui
AU  - Anup Rao
AU  - Jayakumar Subramanian
AU  - Branislav Kveton
PY  - 2026
UR  - https://arxiv.org/abs/2506.06964
ID  - 2506.06964
ER  -