We would like to invite to the first international meeting of AI Section.
The first international guest of AI Section - Panayiotis Panayiotou from University of Bath is going to visit TUL on 10.06.2024. The meeting will start at 6:15 PM and will be held in CTI.
Topic:
A Dive into Reinforcement Learning
This talk will provide a comprehensive overview of the foundational reinforcement learning (RL) methods. We will start by exploring Dynamic Programming (DP), which provides a framework for solving complex problems by breaking them down into simpler subproblems. Following this, we will look into Monte Carlo (MC) methods, which learn directly from episodes of experience without requiring a model of the environment's dynamics. We will then review Temporal Difference (TD) learning, which combines ideas from MC and DP to learn directly from raw experience while still using learned estimates.
Each method will be discussed regarding its algorithmic approach, advantages, and practical applications, providing a clear picture of how to choose the most effective approach for each problem. To conclude, we will briefly touch upon how these foundational methods pave the way for advanced Deep Reinforcement Learning (DRL) techniques and highlight DRL's revolutionary impact on gaming, robotics, and beyond.
Bio:
Panayiotis Panayiotou (Pani) is a PhD student at the University of Bath, working on causal representations in Reinforcement Learning. He worked as a software engineer at Google until January 2024. Previously, he completed Master's Degree at ETH Zürich, and Bachelor's at Imperial College London, both in Computer Science.