Tag: Calibration-Aware Policy Optimization
Browse our exclusive articles!
T2PO: Stable Multi-Turn RL with Uncertainty-Guided Exploration
Discover T2PO, an uncertainty-guided framework that stabilizes multi-turn reinforcement learning by optimizing exploration at token and turn levels.
Enhancing Multi-Hop Reasoning with Structural Causal Models
Discover how Structural Causal Models and Group Relative Policy Optimization improve multi-hop fact verification accuracy and interpretability.
Segment-Aligned Policy Optimization for Multi-Modal AI Reasoning
Discover how Segment-Aligned Policy Optimization (SAPO) enhances multi-modal reasoning in AI by improving policy updates and training stability.
PORTool: Optimizing Multi-Tool AI Reasoning with Rewarded Trees
Discover how PORTool improves AI multi-tool reasoning using importance-aware policy optimization and rewarded rollout trees for better accuracy and efficie...
Safe Reinforcement Learning with Augmented Lagrangian Network
Discover how Augmented Lagrangian Multiplier Network enhances state-wise safety and stability in reinforcement learning for real-world applications.