concept2code: Deep Reinforcement Learning for Conversational AI

concept2code: Deep Reinforcement Learning for Conversational AI

연구 분야: Artificial Intelligence

학회: KDD '22: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Deep Reinforcement Learning uses best of both Reinforcement Learning and Deep Learning for solving problems which cannot be addressed by them individually. Deep Reinforcement Learning has been used widely for games, robotics etc. Limited work has been done for applying Deep Reinforcement Learning for Conversational AI. Hence, in this tutorial cover application of Deep Reinforcement Learning for Conversational AI. We give conceptual introduction to Reinforcement Learning and Deep Reinforcement Learning. We then present various real-life approaches with increasing complexity in detail. The approaches include dialog generation, task-oriented dialog generation, modelling chitchat, natural language generation, hierarchical, weakly supervised, multi-domain and decision transformer. We then walk-through code for implementation of core ideas and for some of the real-life approaches.