SFB 876 - News

Resource-aware Machine Learning - International Summer School 2012

Find information about the summer school 2014 here

Focus of the summer school

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.

Read below about the previous summer school in 2012

From to , the TU Dortmund University, Germany, hosted a summer school addressing solutions to these constraints.

For the summer school, world leading researchers in machine learning and embedded systems gave lectures on several techniques dealing with huge amounts of data, distributed data and constraints of embedded systems.

The summer school was open for international PhD or advanced master students, who wanted to learn cutting edge techniques for machine learning with constrained resources. Student grants have been available to fund travel and accommodation.

The international summer school on resource-aware maching learning was part of the graduate school of the Collaborative Research Center SFB 876. The research center targets both ends of the machine learning spectrum: Small devices, with high restrictions due to the device, and high-dimensional data with a complexity exceeding even large computing center's capacity.


A selection of the topics of the lectures:
  • Mining ubiquitous data streams
  • Practical application of machine learning and statistics tools: RapidMiner and R
  • Massively parallel programming: Use the power of your GPU
  • Being aware of energy constraints: Battery models
  • Which model is best? Criteria for model quality assessment to enhance the selection process
  • Towards Self-Powered Systems: Wireless sensor networks and energy harvesting
  • Dealing with memory hierarchies to improve computation performance
For more details on the lectures see here: Lecture Topics.


Located in the heart of the urban area around the river Ruhr, the TU Dortmund University has a long history in research. 40 years ago, the computer science faculty was founded as one of the first of its kind. Since then, the faculty has grown to more than 2.000 students originating from nearly 50 countries all over the world.

The city of Dortmund managed the transition from a center of coal mining and steel milling to information technology. Several historic sites let visitors discover the past from medieval times to former mining facilities.

The summer school took place at the university's Mechanical Engineering main building, Leonhard-Euler-Str. 5. Have a look at location to find more details.

Data Mining Competition

Part of the summer school was a data mining competition between all participants and on real world mobile data. Participants tested there existing and freshly learned skills on machine learning on unexplored data captured on smartphones.

Data from mobile phones is a perfect example for the task of mining distributed data. Often, users carry the devices with them all day, delivering a fine grained usage profile. Together with location information the data does not only enable a view of when something occurs, but also where. Anonymized examples in parallel protect the privacy of individual users.

The competition enabled the participants to get an impression of what is feasible with current mobile phone's capabilities.