Hi, my name is Michel Tokic and I like to welcome you to my website!
At this place, I share some of my research stuff and publications with the community. My research interests are in artificial intelligence with an emphasis on reinforcement learning. I love to develop artificial agents that learn skills through sensorimotor interactions.
In 2013 I attained my PhD at the Institute of Neural Information Processing @ Ulm University under the supervision of Prof. Günther Palm. The title of my dissertation was „Reinforcement Learning with Adaptive Control of Exploration and Exploitation“. In 2014 I joined Siemens AG / Technology in Munich as a research scientist, where my current focus is on data analytics and artificial intelligence for production machines, traffic systems and gasturbines. During the past years I trained neural control policies for production machines and gasturbines, and modelled high-dimensional dynamical systems using recurrent neural networks.
Press & News:
- AI Trilogy Part 3: Learned after 5000 crashes, https://www.siemens.com/global/en/company/stories/research-technologies/artificial-intelligence/ai-trilogy-part2-autonomous-systems.html (German version)
- Industrial AI Podcast „Industrial AI and Motion Control – or how do we reduce engineering time?“, https://aipod.de/podcast-archive/183
- Siemens News: „AI and the choclate factory,“ https://www.siemens.com/global/en/company/stories/research-technologies/artificial-intelligence/ai-and-the-chocolate-factory.html
- Siemens Ingenuity: „AI makes cities livable by forecasting traffic light phases“, https://ingenuity.siemens.com/2021/11/ai-makes-cities-livable-by-forecasting-traffic-light-phases/