XJTU AI Research is dedicated to researching cutting-edge AI technologies. We firmly believe that the value of technology lies in its ability to enhance productivity, rather than solely in being published in high-impact journals or conferences—at least, this should not be the only criterion for evaluation. Our focus areas include

• Medical AI
• Green AI
• Material AI
• Art AI
• Large Language Models
• Embodied Intelligence
• Swarm Intelligence
• Multi-agent Reinforcement Learning
  
  Our goal is to push the boundaries of artificial intelligence and develop innovative solutions that address complex challenges across various domains.

HistoGym

Is the diagnostic process of doctors better modeled as a decision-making task rather than a classification task? To address this question, we propose HistoGym, a custom gym environment for Whole Slide Image (WSI) analysis.