Lecture
Mini-Batches in On- and Off-Policy Deep Reinforcement Learning
In course
DEMO: Lorem in id
Nisi exercitation nulla voluptate et velit duis. Excepteur nostrud consectetur pariatur ex adipisicing duis. Eu id in ex excepteur cillum. Sit voluptate anim consectetur proident.
Description
This lecture covers the importance of mini-batches in Deep Reinforcement Learning, explaining how to avoid data correlation by using replay buffers or multiple actors. It discusses on-policy and off-policy methods, such as Q-Learning and Advantage Actor-Critic, and the pros and cons of each approach.
Instructor
ea labore proident nisi
Nostrud aliqua officia elit velit pariatur cupidatat ex proident Lorem quis adipisicing officia exercitation. Enim laborum commodo commodo veniam mollit minim id consectetur. Veniam magna enim velit amet elit cupidatat ad magna adipisicing dolor. Et pariatur qui excepteur aute irure aliqua. Fugiat dolor ad occaecat et minim magna duis eiusmod consectetur. Nisi ea minim excepteur nisi enim aliqua. Occaecat nostrud mollit enim aliqua sint dolore magna laboris.
Official source
About this result
This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.
Related lectures (37)
Reinforcement Learning: Basics and Applications
Covers the basics of reinforcement learning, including trial-and-error learning, Q-learning, deep RL, and applications in gaming and planning.
Deep Learning Agents: Reinforcement Learning
Explores Deep Learning Agents in Reinforcement Learning, emphasizing neural network approximations and challenges in training multiagent systems.
Learning Agents: Exploration-Exploitation Tradeoff
Explores the exploration-exploitation tradeoff in learning unknown effects of actions using multi-armed bandits and Q-learning.
Reinforcement Learning: Basics and Applications
Covers the basics of reinforcement learning, including Markov Decision Processes and policy gradient methods, and explores real-world applications and recent advances.
Reinforcement Learning: Q-Learning
Covers Q-Learning in reinforcement learning, exploring action values, policies, and the societal impact of algorithms.