Yao Ma

This paper explore the essence of message passing on heterophilic graphs from the perspective of model and design a method with the guidance of theoretical findings.

We propose a lesson-based framework for multiple LLM agents to collaboratively solve coding problems by learning from each other.