Learning Complex Tasks

Chelsea and Daniel discuss a paper that tackles the challenge of scaling model-based RL to solve complex tasks involving multiple segments. The paper introduces a method that combines learned models with value functions to solve more challenging tasks. The results show that the robot is able to learn and sequence primitive skills to perform longer horizon behaviors, demonstrating promising potential for real-world applications.

In this clip
From this podcast
The Gradient
Chelsea Finn on Meta Learning & Model Based Reinforcement Learning
Related Questions
- How is machine learning used in robots as discussed in the episode Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10 and the clip Hierarchical Reasoning Challenges?
- What is the main topic of the clip Learning Robotics, Meta Learning from the episode Chelsea Finn on Meta Learning & Model Based Reinforcement Learning?

Learning Complex Tasks

In this clip

From this podcast

The Gradient

Chelsea Finn on Meta Learning & Model Based Reinforcement Learning

Related Questions

How is machine learning used in robots as discussed in the episode Pieter Abbeel: Deep Reinforcement Learning | Lex Fridman Podcast #10 and the clip Hierarchical Reasoning Challenges?

What is the main topic of the clip Learning Robotics, Meta Learning from the episode Chelsea Finn on Meta Learning & Model Based Reinforcement Learning?