EvonNemo - A Symbiosis of Datalog Tracing and Proof Tree Visualization

Lukas GerlachLukas Gerlach, Alex IvlievAlex Ivliev, Julián MéndezJulián Méndez, Simon MeuselSimon Meusel, Raimund DachseltRaimund Dachselt, Markus KrötzschMarkus Krötzsch

Download

Lukas Gerlach, Alex Ivliev, Julián Méndez, Simon Meusel, Raimund Dachselt, Markus Krötzsch
EvonNemo - A Symbiosis of Datalog Tracing and Proof Tree Visualization
The Fifth Workshop on Explainable Logic-Based Knowledge Representation (XLoKR 2024), November 2024

Details
Bibtex

KurzfassungAbstract
Datalog and other rule-based formalisms are explainable by design, but showing explanations in a digestible form to a human remains a key challenge since rule engines often work with billions of facts. As a first step in this direction, we present a prototype of the integration of two KR tools: Nemo, a fast and versatile rule reasoning toolkit, and Evonne, a tool for visualizing, analysing, and debugging Description Logic proofs. We briefly discuss challenges on both ends such as computing Datalog proof trees in Nemo and special requirements for their visualization concerning Evonne. We conclude this extended abstract by showcasing key analysis features and an elaborate discussion of our roadmap to further aid explainability as well as developer experience from both a knowledge engineering perspective and a rule engine point of view.
Weitere Informationen unter:Further Information: Link
Projekt:Project: Cfaed, CPEC, InnoSale, SECAI, ScaDS.AI
Verknüpfte Tools:Related Tools: Nemo
Forschungsgruppe:Research Group: Wissensbasierte SystemeKnowledge-Based Systems

@inproceedings{GIMMDK2024,
  author    = {Lukas Gerlach and Alex Ivliev and Juli{\'{a}}n M{\'{e}}ndez and
               Simon Meusel and Raimund Dachselt and Markus Kr{\"{o}}tzsch},
  title     = {EvonNemo - A Symbiosis of Datalog Tracing and Proof Tree
               Visualization},
  booktitle = {The Fifth Workshop on Explainable Logic-Based Knowledge
               Representation (XLoKR 2024)},
  year      = {2024},
  month     = {November}
}