Project description

Graph transformation is a well-established theory that studies computational methods to transform graphs in a stepwise manner by the application of simple rules which are discrete in nature. This project under the leadership of Professor Frank Drewes aims to integrate neural methods of computation, and thus continuous aspects, into rule-based models of graph transformation in order to combine the individual strengths of both paradigms. Rule-based models are transparent and explainable; they make sense to humans and are accessible to algorithmic techniques while neural models are adaptive and learnable. The aim of this project is to develop models which combine these advantages. The project includes both formalization and mathematical reasoning on the one hand, and implementation for the purpose of experiments and demonstration on the other hand.

Admission requirements

The general admission requirements for doctoral studies...