Montreal Robotics Courses

Safe and Deployable Reinforcement Learning for Reason and Action

Explore reinforcement learning applications in robotics, building confidence in AI agents, and training LLMs to reason and act simultaneously for real-world deployment.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

Safe and Deployable Reinforcement Learning for Reason and Action

Explore reinforcement learning challenges in robotics, building AI agent confidence, and training LLMs to reason and act simultaneously in real-world applications.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

RobotLearning: Scaling Offline Reinforcement Learning

Explore offline reinforcement learning techniques for robots, including trajectory stitching, model-based approaches, Decision Transformers, and methods for integrating offline data into training pipelines.

• YouTube
• 1 hour 47 minutes
• Self-Paced
• Free Video

Hierarchical Planning with Large Models for Robot Learning

Explore hierarchical planning and reinforcement learning for robotics, from humanoid control fundamentals to advanced applications with large models and policy co-training techniques.

• YouTube
• 1 hour 51 minutes
• Self-Paced
• Free Video

Robot Learning: Multi-Agent Reinforcement Learning and RLHF

Explore the foundations of multi-agent reinforcement learning (MARL) and its connection to reinforcement learning from human feedback (RLHF), including structured learning, algorithms, and training large language models.

• YouTube
• 1 hour 47 minutes
• Self-Paced
• Free Video

Gemini Robotics: Robot Learning and Capabilities

Explore the latest advancements in Gemini Robotics through Keerthana Gopalakrishnan's presentation on recent developments and findings in robot learning technology.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Scaling Deep Q-Learning - Part 2

Delve into the challenges of deep Q-learning, exploring target networks, double Q-learning, and the "deadly triad" problem, with applications in robotics through QT-Opt and PQ-N algorithms.

• YouTube
• 1 hour 18 minutes
• Self-Paced
• Free Video

Scaling Deep Q-Learning - Part 1

Explore the progression from multi-armed bandits to Q-learning, examining exploration-exploitation challenges, approximate dynamic programming, and efficient policy iteration methods in reinforcement learning.

• YouTube
• 37 minutes
• Self-Paced
• Free Video

Scaling Continuous Deep QLearning - Part 2

Delve into deterministic policy gradient methods, focusing on DDPG for continuous action spaces, target networks, exploration strategies, and recent advancements in deep Q-learning scalability.

• YouTube
• 1 hour 43 minutes
• Self-Paced
• Free Video

Scaling Continuous Deep QLearning - Part 1

Delve into deterministic policy gradient methods, focusing on DDPG for continuous action spaces. Learn about target networks, policy updates, exploration noise, and recent advancements to improve robustness and scalability in deep reinforcement learning.

• YouTube
• 13 minutes
• Self-Paced
• Free Video

Scaling Policy Gradients for Reinforcement Learning in Robotics - Part 1

Dive into reinforcement learning for robotics with a focus on policy gradients, their mathematical foundations, and practical applications in autonomous systems, including a guided homework assignment.

• YouTube
• 1 hour 49 minutes
• Self-Paced
• Free Video

Scaling Policy Gradients Part 2

Delve into variance reduction techniques for policy gradients, exploring reward-to-go, critics, and the bias-variance trade-off, with practical insights from AlphaStar's implementation in StarCraft.

• YouTube
• 57 minutes
• Self-Paced
• Free Video

Robot Learning: Methods and Considerations for Scaling Data Collection

Explore the challenges and methodologies of robotic data collection, from traditional programming approaches to modern solutions integrating expert demonstrations and third-person video datasets.

• YouTube
• 1 hour 45 minutes
• Self-Paced
• Free Video

Learning Reward Models and Using Foundational Models in Robot Learning

Explore reward functions in reinforcement learning, from basic concepts to advanced applications using foundational models, with insights on intrinsic and extrinsic rewards in real-world robotics.

• YouTube
• 1 hour 36 minutes
• Self-Paced
• Free Video

Robot Learning: Learning to Plan with Large Models - Part 2

Delve into latent inverse models and their applications in robotics, exploring how AI predicts future events through world dynamics, action control, and state inference from visual data.

• YouTube
• 49 minutes
• Self-Paced
• Free Video