Xi-Qian Lu

Large Language Models and Reinforcement Learning: A Taxonomy of Integration Paradigms, Challenges, and Future Directions

This paper highlights the transition from static prediction to sequential decision-making, emphasizing RL’s strengths in long-term reward optimization and interaction modeling, and LLMs’ advantages in semantic understanding and reasoning.