Research spans generative and multimodal models, LLM‑based agents, and reinforcement learning for goal‑oriented decision‑making. Applies AI to personalized learning and computational social simulation, emphasizing rigorous evaluation, scalability, and practical deployment.