392286 ISY Project: Learning to Act - Reinforcement Learning with Deep Neural Networks in Simulated Environments - group 2 (Pj) (SoSe 2018)

- Short Description
How can a system learn to act and achieve rewards from experience? In this project, you will apply neural networks to learn and produce complex purposeful actions in simulated environments. During the last years, such methods became successful in autonomous learning of winning strategies in computer game scenarios. You will apply Deep Reinforcement Learning in the context of Atari games (1) and for solving locomotion (2):
(1) As an example for learning in the context of games, see: OpenAI Gym (https://github.com/openai/gym) and for an implementation of a RL approach http://karpathy.github.io/2016/05/31/rl/
(2) In the locomotion task you will develop a learning controller to enable a physiologically-based human model to navigate a complex obstacle course as quickly as possible. You are provided with a human musculoskeletal model and a physics-based simulation environment where you can synthesize physically and physiologically accurate motion.

Please note that the teams will be selected by the supervisors on the basis of short applications that students are expected to send to them. Registering to the project in the ekVV will only be regarded as expression of interest; it will not secure a team membership.
Please get in touch with the supervisors for information on the application procedure.

Requirements for participation, required level

Required skills:
- Introduction to Neural Networks course (or Advanced Neural Networks)
- Python (>= 1 year)