Predictor function (P01)

• Subject Area: Theoretical Computer Science
• Term: from 2016 to 2020
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 276397488

The predictive function is a central part of the CRC1232, as it maps micro-descriptors to material properties, using macro descriptors as an intermediary. This project aims to significantly extend this predictive function by incorporating annotated explanations to achieve transparency in its predictions in order to discover formerly unknown correlations. To allow efficient reasoning about the high dimensional process-parameter space, synthetic descriptors will be determined by applying verification techniques from boolean algebra to the complex models underlying both mechanical and thermal treatment processes. Finally, the development of a heuristic microprocessfunction will help determine promising candidates for materials with given properties.

• DFG Programme: Collaborative Research Centres
• Subproject of: SFB 1232: From colored states to evolutionary structural materials
• Applicant Institution: Universität Bremen
• Project Head: Professor Dr. Rolf Drechsler