YoVDO

MOPO - Model-Based Offline Policy Optimization

Offered By: Simons Institute via YouTube

Tags

Deep Reinforcement Learning Courses

Course Description

Overview

Explore a deep reinforcement learning presentation on Model-Based Offline Policy Optimization (MOPO) delivered by Tengyu Ma from Stanford University at the Simons Institute. Delve into topics such as distributional domain shift, answer identification, and improved sketch techniques as the speaker discusses innovative approaches to offline reinforcement learning. Gain insights into the challenges and solutions in developing effective policies from pre-collected datasets without direct interaction with the environment.

Syllabus

Introduction
Workshop Overview
Presentation
Distributional Domain Shift
Answer Identification
Improved Sketch
Summary
Discussion


Taught by

Simons Institute

Related Courses

6.S094: Deep Learning for Self-Driving Cars
Massachusetts Institute of Technology via Independent
Natural Language Processing (NLP)
Microsoft via edX
Deep Reinforcement Learning
Nvidia Deep Learning Institute via Udacity
Advanced AI: Deep Reinforcement Learning in Python
Udemy
Self-driving go-kart with Unity-ML
Udemy