Michael Somma

Research Associate, JOANNEUM RESEARCH

As an AI Researcher at JOANNEUM RESEARCH in Graz, Michael brings a broad portfolio in artificial intelligence, having worked extensively in both academic and commercial settings. As a Research Associate at the University of Vienna, he developed innovative computer vision models that contributed significantly to two FWF-funded research projects. At Deloitte, he led the deployment of advanced natural language processing solutions and was instrumental in pioneering reinforcement learning applications. His expertise spans developing automated trading systems with Deep Reinforcement Learning and integrating Deep Reinforcement Learning into physical simulations.



Deep reinforcement learning


This workshop offers a comprehensive exploration of Reinforcement Learning (RL), starting with a brief recap of the fundamental concepts. We will then advance to Deep Reinforcement Learning, focusing on the Deep Q-Learning Algorithm. Participants will engage in a hands-on lab featuring two practical examples: the lunar lander and the cart pole. The session will cover setting up the RL environment, implementing the agent, and training the agents from scratch. We will also compare the performance of these trained agents against untrained random agents. All activities will be supported by a Google Colab Notebook, provided to facilitate real-time coding and experimentation.

Main Partners


Other Partners