conceptual_blending_project

Scientific team projects for master students of the Faculty of Computer Science (FIN).

Computational Conceptual Blending: a team of 4-6 students lead by Mihai Codescu, Fabian Neuhaus, Madhura Thosar

The goal of the project is to develop a simplified system which blends two or more ontologies about familiar concepts to create a new concept. The idea is to use the system to create monsters by blending two or more ontologies about animals. The system can be roughly divided into three modules, namely; extraction, blend generation and evaluation. The task to be carried out in order to develop each module are listed below.

The following tasks are to be carried out in this project.

1. Literature search on blending patterns
2. Create blending patterns repository
3. Select the blending pattern from a given list of patterns with respect to requirements and background knowledge
4. Select the input spaces according to the blending pattern and blending requirement
5. Generate a base and base morphism from input spaces using structure mapping and anti-unification techniques
6. Generate the blendoid from the input spaces using base and base morphism
7. Evaluate the blendoid for consistency. If the blendoid is inconsistent, then weaken the input spaces using conflict sets until the blendoid is consistent.
8. Evalute the consistent blendoid to check if it satisfies the requirements.