Talk: From Deep Reinforcement Learning to LLM-based Agents: Perspectives on Current Research