Overview
Learn about the fundamental concepts of planning in sequential decision-making problems through a comprehensive lecture that explores both theoretical frameworks and practical applications in robotics. Dive into the process of finding optimal actions that lead to desired states, understanding how agents discover solutions in real-world scenarios where perfect solutions may not be immediately apparent. Explore the computational challenges in planning, particularly in stochastic environments, while examining crucial factors such as branching factors and search tree depth that impact solution optimization. Master the identification and exploitation of structural properties including compositionality, productivity, and localism to enhance the efficiency of planning algorithms and overcome computational hurdles in robotic applications.
Syllabus
Robot Learning: Sequential Decision Making
Taught by
Montreal Robotics