Joint reinforcement learning and self-distillation combines task rewards with dense teacher guidance for agent training, but their gradients can partly cancel. A mixing coefficient controls relative scale without selectively removing opposing components. We present CORD, a general framework for conflict-aware optimizat...
Wen-Hao Yu, Zheng-Cheng Wang, Jia-Hong Liu et al.· 0 citations
A Graph-based Faithful sTep-level credit-assignment framework (GRAFT) that grafts all rollout trajectories into a trajectory graph, recovering node state-values via Bellman iteration on the graph, and assigning credit to each edge by the node value difference is proposed.
Xin-Cheng Yao, Haobo Fu, Wei-Ming Liu et al.· arXiv.org· 0 citations
AutoPref is presented, the first LLM-guided framework for automated preference-objective discovery in NCO, which consistently outperforms strong hand-designed baselines across problem scales, demonstrating the benefits and scalability of automated objective discovery for NCO.
Shengda Gu, Kai Li, Xin-Yi Ke et al.· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.