Deep Reinforcement Learning for Sequential Decision Making Tasks with Natural Language Interaction
Center for Language & Speech Processing(CLSP), JHU via YouTube
Overview
Explore deep reinforcement learning applications for sequential decision-making tasks involving natural language in this lecture by Dr. Satinder Singh from the University of Michigan. Delve into two projects that combine deep learning advances in visual perception and natural language processing with reinforcement learning. Examine approaches for querying, reasoning, and answering questions on ambiguous texts relevant to dialog systems. Investigate a hierarchical deep reinforcement learning architecture enabling zero-shot generalization to unseen instructions in 3D maze navigation. Gain insights into cutting-edge research bridging natural language understanding and reinforcement learning for complex interactive tasks.
Syllabus
Deep Reinforcement Learning for Sequential Decision Making Tasks with Natural Language Interaction
Taught by
Center for Language & Speech Processing(CLSP), JHU