Explore a comprehensive video analysis of a research paper on offline reinforcement learning benchmarks. Delve into the challenges of evaluating offline RL algorithms and learn about a new benchmark designed to address these issues. Discover key properties of datasets relevant to offline RL applications, including those generated by hand-designed controllers and human demonstrators, multi-objective datasets, and heterogeneous mixes of trajectory quality. Understand how this benchmark aims to focus research efforts on methods that can drive substantial improvements in real-world offline RL problems. Gain insights into the paper's abstract, authors, and access links to the full paper and associated code repository.
Datasets for Data-Driven Reinforcement Learning
Overview
Syllabus
Datasets for Data-Driven Reinforcement Learning
Taught by
Yannic Kilcher
Related Courses
-
Understanding Algorithms for Reinforcement Learning
-
Reinforcement Learning for Trading Strategies
-
Reinforcement Learning with Gymnasium in Python
-
Reinforcement Learning
2.9 -
AI skills: Introduction to Unsupervised, Deep and Reinforcement Learning
-
Unsupervised Learning, Recommenders, Reinforcement Learning
