Piter Py #4

 

Reinforcement learning is about developing strategies to solve decision-making problems. This field has been dominated for researchers and people with an academical background, and some of the concepts may appear confusing to people coming from different areas.
This talk will be about introducing reinforcement learning from a developer’s perspective. We will start explaining the core concepts of the discipline with no mathematical formulation involved, showing the latest achievements in this field, and providing a list of useful resources about the topic.
Then, we will talk about the infrastructure required to solve an RL problem, what are the different open source tools at our disposal, and what are best practices when implementing RL algorithms.
Finally, we will show a practical example of how to code, train, and use an agent capable of solving an Atari Game (MsPacman) with an state of the art algorithm.

 

Spain. Madrid

AI consultant

 

Guillem Duran is a telecommunications engineer who is passionate about science, computers, and artificial intelligence. He is a member of the PyData Mallorca organization, and he loves sharing knowledge and code at talks, workshops and courses.
Although he started learning about Machine Learning for the pleasure of finding things out, during the last five years he has been collaborating as a data science consultant for several companies and universities. Now he is working at source{d}, developing an open source framework for reinforcement learning research.