Block-Course (around end of semester)
Elective course for M.Sc. CIT, Electrical Engineering & Information Technology and Space-ST.
Contents:
• Why swarm exploration – some motivational examples with applications in space and on Earth
• Preliminaries: selected mathematical formalism (vector spaces, matrices, representation and approximations in vector spaces, least squares, convex optimization)
• Recap of probability and statistics (calculus of probabilities, moments, Bayesian theory)
• Machine learning tools (supervised learning, linear regression, kernel methods, neural networks, impact of regularization, sparsity and compressed sensing)
• Models for multi-agent networks (connected network models, distributed inference strategies)
• Distributed machine learning and exploration (models for static spatial regression, information-theoretic exploration approaches, Bayesian sequential methods for learning and exploration)
• Discussion of several practical examples of swarm exploration solutions (cooperative localization, information-driven sparse mapping of magnetic fields, exploration of sparse gas sources using a swarm of mobile robots)
As outcome, the students should be able to:
• Understand key concepts in distributed information processing over networks,
• Explain and apply mathematical tools needed to implement classical machine learning algorithms in distributed settings
