Methods for Efficient Resource Utilization in Machine Learning Algorithms (A03)

• Subject Area: Computer Architecture, Embedded and Massively Parallel Systems ; Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: from 2011 to 2022
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 124020371

This subproject builds a bridge between learning algorithms and resource efficiency. During the second phase of this subproject, we intend to develop methods for automatic model selection in machine learning, which are making efficient use of available resources. We want to select from a large number of compute-intensive learning techniques the ones exhibiting the best predictions, for problems incorporating a large number of observations or variables. Using personalized medicine for demonstration, we want to show how jointly extending methods of machine learning and real-time systems allows us to solve problems, which are currently exceeding the state of the art in predicting the impact of medication.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 876: Providing Information by Resource-Constrained Data Analysis
• Applicant Institution: Technische Universität Dortmund
• Project Heads: Professor Dr. Jian-Jia Chen, since 1/2019; Professor Dr. Peter Marwedel, until 12/2018; Professor Dr. Jörg Rahnenführer