Resource-aware Machine Learning

About

Machine learning is the key technology to discover information and concepts hidden in huge amounts of data. At the same time, availability of data is ever increasing. Better sensors deliver more accurate and fine-grained data, more sensors a more complete view of the scenario. While this should lead to better learning results, it comes at a cost: Resources for the learning task are limited, restricted by computational power, communication restrictions or energy constraints. The increased complexity needs a new class of algorithms respecting the constraints.

Course leader

Rakesh Agrawal, Sebastian Buschjäger, Kristian Kersting, Thomas Liebig, Wayne Luk, Mojtaba Masoudinejad, Rob Maunder, Katharina Morik, Nico Piatkowski, Chris Schwiegelshohn, Olaf Spinczyk

Target group

PhD or advanced master students from computer science or related disciplines using machine learning as an application (e.g. astrophysics, biology, medicine, engineering)

Course aim

Topics of the lectures include: Machine learning on FPGAs, Deep Learning, Probabilistic Graphical Models and Ultra Low Power Learning.

Exercises help bringing the contents of the lectures to life. The PhyNode low power computation platform was developed at the collaborative research center SFB 876. It enables sensing and machine learning for transport and logistic scenarios. These devices provide the background for hands-on experiments with the nodes in the freshly built logistics test lab. Solve prediction tasks under very constrained resources and balance accuracy versus energy.

Fee info