Marco Huber: Reinforcement Learning – How Machines learn like Humans do
22. June 2017
Humans and animals learn by interacting with their environment. They react and adapt their behavior depending on the perceived consequences. In doing so, causes and effects are processed directly, without utilizing labeled data or instructions provided by an expert. This natural learning paradigm is exploited in a special branch of machine learning called Reinforcement Learning (RL). RL has been successfully applied in past for instance in teaching computers to play board games like Backgammon or Go. Robotics is another domain, where RL is crucial for teaching complex tasks like tossing a pancake or catching a ball. In this talk, an introduction to the major streams and basic algorithms of RL is given. Furthermore, the link between RL and biological learning is highlighted.
The Talk of this Meetup his held by Marco Huber of USU AG.