Dynamic Policy-Driven Adaptive Multi-Instance Learning for Whole Slide Image Classification
The author proposes a Dynamic Policy-Driven Adaptive Multi-Instance Learning (PAMIL) method for Whole Slide Image (WSI) tasks, integrating dynamic instance sampling and reinforcement learning to improve feature selection and decision-making.