This paper introduces UI-MOPD, a continual-learning method for GUI agents that uses multi-teacher on-policy distillation across platforms. It also presents Uni-GUI, a cross-platform GUI interaction dataset designed to address limited platform coverage and scarce executable trajectories. The method selects a platform-specific teacher based on the current environment and transfers behavioral priors into a shared policy through platform-conditioned distillation. On OSWorld and MobileWorld, UI-MOPD reports task success rates of 38.2% and 12.0%, respectively. The work targets capability retention on existing platforms while adapting to new ones.
No heat snapshots are available in the last 24 hours.