Bayesian Action Evaluation
Kamyar discusses how he evaluates actions using a distribution that represents both expected value and uncertainty. By sampling from this distribution, he can quantify his beliefs about the goodness of each action, illustrating the concept of Bayesian inference in decision-making. The conversation highlights the importance of understanding variance and how it influences perceived value, drawing a parallel between personal beliefs and statistical distributions.In this clip
From this podcast
The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence)
Deep Reinforcement Learning Primer and Research Frontiers with Kamyar Azizzadenesheli - TWiML...
Related Questions