Catalog description: 

Learning and development in humans. Studies in developmental psychology and cognitive science. Novelty, curiosity, and surprise. Social interaction as a means for learning and development. Perception, learning and use of affordances in humans. Language development in children. Robotics as a test-bed for ideas developed in psychology and cognitive science. Review of recent studies in developmental robotics.

Machine learning in robotics. Supervised, unsupervised and reinforcement learning. Markov decision process framework to formulate and study learning problems in robotics. Challenges that robotics poses for using reinforcement learning methods. Inverse reinforcement learning. Apprenticeship learning. Learning from demonstration. 

Course description: The course will focus on the study of learning and development methods to robots to enable them to acquire new skills and/or adapt to the changes in their environment. In this sense, it is positioned in the junction of robotics, machine learning and cognitive science (and psychology). Robotics is a challenging test-bed for learning and development, and often uncovers implicit assumptions and shortcomings of learning methods that are developed in more theoretical and abstract domains. The course will briefly introduce the main tracks of research in learning and development and will focus on how these methods can be applied to problems on robots. The course will appeal to students from Computer Engineering as well as students from Electrical and Electronics, and Mechanical Engineering, and Cognitive Science (Informatics Institute). 

Course Objectives:

At the end of the course, the student will be able to:

-Discuss the challenges that robotics poses to learning and development methods

-Describe and formulate robotics problems in machine learning frameworks

-Read and understand literature on the use of learning and development methodologies in robotics

-Implement and analyze a learning/development method on a simulated robotic problem