Interview with Hikaru Shindo and Quentin Delfosse: Neurosymbolic Reinforcement Learning with NUDGE скачать в хорошем качестве

Interview with Hikaru Shindo and Quentin Delfosse: Neurosymbolic Reinforcement Learning with NUDGE

Really great conversation with Hikaru Shindo and Quentin Defosse on their innovative new approach to RL called NUDGE that allows for the learning of policies that are rule-based using a neural approach. You can find the paper here: https://arxiv.org/abs/2306.01439 0:00:00 Introduction 0:08:00 AI Mis-alignment 0:16:48 Nudge 0:20:48 Integrating differntiable reasoning into RL 0:33:53 Neurosymbolic policy learning 0:40:44 Integrating object discovery into neurosymbolic RL 0:48:04 Experiments 0:52:20 Coffee break (it was early in Arizona when we filmed) 0:52:33 Back to experiments 1:05:20 Explainability of policies 1:10:06 Conclusion The Neuro Symbolic Channel provides the tutorials, courses, and research results on one of the most exciting areas in artificial intelligence and machine learning. With content originally from the AI course taught at Arizona State University, this channel brings you the latest at the intersection of symbolic methods (e.g., logic programming) and deep learning. Learn about the latest algorithms, Python packages, and progress toward larger goals such as artificial general intelligence (AGI).