The Learnable Universe скачать в хорошем качестве

The Learnable Universe

Treating AI as a philosophical project by Joscha Bach. Why do we find ourselves in a universe that has learnable properties? How is it possible for a symbol to mean something? What is the relationship between observation, perception and knowledge? What is agency? What constitutes a self model? When we approach Artificial Intelligence as a philosophical project, we gain a fascinating and useful perspective on age-old questions of philosophy. This short presentation will touch on some of these questions and aims to open up a broader space for discussion. Our speaker, Joscha Bach, PhD, is a cognitive scientist and AI researcher specializing in computational models of cognition and neuro-symbolic AI. He has taught and worked in AI research at Humboldt University of Berlin, the Institute for Cognitive Science in Osnabrück, the MIT Media Lab, the Harvard Program for Evolutionary Dynamics. Currently, he serves as a principal AI scientist at Intel Labs, California. Chapters: 0:00 - Introduction 2:42 - AI as a philosophical project: Philosophy, Mathematics and Computation 11:20 - Symbolic AI, Deep Learning, Meta Learning, Artificial Neurons, Neural Circuitry in ANNs 16:57 - Free Energy Principle 19:25 - Cybernetics: Modeling in the service of control. Part 1 21:13 - Good Regulator Theorem 22:06 - The Controllable Universe 24:00 - Is the Universe a Computer? 30:55 - Cybernetics: Modeling in the service of control. Part 2 36:00 - The Four Types of Representation Anchors: possibility, probability, valence, normativity 36:20 - Generalization of Model 36:54 - Assimilation and Accommodation 37:45 - Coherence 38:20 - Attention Agent, Motivation Agent, Perception Agent 41:00 - Q&A Join Buzzrobot Slack community to connect with fellow data scientists, ML engineers and researchers: https://buzzrobot.slack.com/join/shar...