Haoyuan Hu

MAPO: Momentum-Aware Policy Optimization

Auto-Rubric: Learning to Extract Generalizable Criteria for Reward Modeling

StepHint: Multi-level Stepwise Hints Enhance Reinforcement Learning to Reason

G2PTL: A Pre-trained Model for Delivery Address and its Applications in Logistics System

Effective Generation of Feasible Solutions for Integer Programming via Guided Diffusion

LaDe: The First Comprehensive Last-mile Express Dataset from Industry