Scene Modification with MR-CKR and ASP
Implementation prototype for scene generation in Autonomous Driving with Multi-Relational CKRs and ASP with Algebraic Measures
This repository provides a prototype for a tool for scene modification using multi-relational Contextual Knowledge Repositories in OWL and ASP with Algebraic Measures, encoded as ASP with weak constraints.
The linked repository includes Python code for building the MR-CKR representation of different types of diagnosis/danger cases and the additional code for ASP encoding of the scene modification process.
Additional Information:
The prototype is based on Python and uses the Java command-line tool CKRew to build the ASP translation.
The resulting ASP code can be used with the clingo ASP solver.
External links:
The code of the prototype is publicly available at
https://github.com/raki123/MR-CKR
including instructions on installation and examples of use.
More information on the related CKRew tool can be found at
References:
The formal description of the scene modification process implemented by the tool is provided in the technical report:
- Loris Bozzato, Thomas Eiter, Rafael Kiesel and Daria Stepanova (2023).
Contextual Reasoning for Scene Generation (Technical Report).
https://arxiv.org/abs/2305.02255
A formal definition of the use of multi-relational CKRs with ASP and Algebraic Measures is given in the paper:
- Loris Bozzato, Thomas Eiter and Rafael Kiesel (2021).
Reasoning on Multi-Relational Contextual Hierarchies via Answer Set Programming with Algebraic Measures.
In 37th International Conference on Logic Programming (ICLP 2021).https://doi.org/10.1017/S1471068421000284