Tuesday, March 19

Curso: Reinforcement Learning – Arizona State University

Dimitri Bertsekas ha anunciado que la serie de videos de su curso Reinforcement Learning (2022) de la universidad Arizona State University, así como también las diapositivas, los apuntes y el PDF del libro “Lessons from AlphaZero for Optimal, Model Predictive, and Adaptive Control” ya son de acceso público y gratuito.

Por Homer Díaz

El curso de Reinforcement Learning es impartido por Dimitri Bertsekas, autor de una variedad de libros y catedrático de las universidades Massachusetts Institute of Technology y Arizona State University (No. 1 -y por siete años consecutivos- en el ranking “Most Innovative Schools” en los EE.UU).

Son 13 videos de aproximadamente 2 horas cada uno y se abordan los siguientes temas:

  • Introduction to exact and approximate dynamic programming
  • Approximation in value and policy space
  • Off-line training, on-line play, and Newton’s method
  • Rollout, approximate policy iteration, and Newton’s method
  • Model predictive and adaptive control
  • Multiagent and multiprocessor reinforcement learning
  • Training of feature-based approximation architectures and neural networks
  • Policy networks and approximation in policy space
  • Aggregation and other problem approximation architectures
  • Applications in engineering, artificial intelligence, and discrete optimization

Puedes acceder a todo el material del curso a través del website: [Website] [Playlist] [Book]

Leave a Reply

Your email address will not be published. Required fields are marked *