Guy Paterson-Jones
Guy Paterson-Jones is a computer scientist whose research centers on defeasible reasoning, the branch of AI logic that handles conditionals of the form “if A, then typically B”. In 2020, they published the Boolean KLM logic (BKLM), shown to be more expressive than the earlier PTL logic, and presented a BKLM paper with Giovanni Casini and Thomas Meyer at the NMR 2020 workshop. As an MSc student at CAIR, they extended the KLM approach to a restricted version of first-order logic in a 2021 paper with the same co-authors. A 2022 master's thesis, supervised by Meyer and Casini, then investigated KLM-style semantics for defeasible reasoning over Datalog knowledge bases.
Timeline
-
September 2020
The BKLM logic at NMR
In September 2020, they present “BKLM - An expressive logic for defeasible reasoning”, written with Giovanni Casini and Thomas Meyer, in a non-monotonic logics session of the virtual NMR 2020 workshop.
-
2020
Boolean KLM at SACAIR
In 2020, they publish “A Boolean Extension of KLM-Style Conditional Reasoning” at the First Southern African Conference for AI Research (SACAIR 2020) in Muldersdrift, South Africa. The paper introduces Boolean KLM (BKLM), a logic that allows disjunctions, conjunctions and negations of defeasible implications without nesting, and shows it is more expressive than PTL.
-
2021
Defeasible reasoning for Datalog
In 2021, they show the project “Defeasible Reasoning for Datalog” at the Project Showcase of UCT's School of Information Technology. The project aims at integrating rules with exceptions into Datalog ontologies.
-
2021
KLM for first-order logic
In 2021, as an MSc student at CAIR, they co-author “KLM-Style Defeasibility for Restricted First-Order Logic” with Giovanni Casini and Thomas Meyer for the 19th International Workshop on Non-Monotonic Reasoning. The paper extends the KLM approach to a restricted version of first-order logic, giving rationality postulates, a semantics and a defeasible entailment inspired by Rational Closure, as a first step toward KLM-style reasoning in Datalog+/- ontology languages.
-
2022
A master's thesis on Datalog
In 2022, they complete the master's thesis “KLM-Style Defeasible Reasoning for Datalog” in the Department of Computer Science, supervised by Thomas Meyer and Giovanni Casini. It builds on the KLM framework, which describes an agent's defeasible knowledge as conditionals of the form “if A, then typically B”, and investigates KLM-style semantics for defeasible reasoning over Datalog knowledge bases.
Sources and credits
Photo: We have no published photo of Guy Paterson-Jones yet. Send us a free-licensed photo.
- NMR 2020 Virtual Program
- Knowledge Representation and Reasoning Research Publications | CAIR
- Defeasible Reasoning for Datalog - Project Showcase 2021
- Guy Paterson-Jones | CAIR
- KLM-Style Defeasibility for Restricted First-Order Logic - UCT Computer Science Research Document Archive
- KLM-Style Defeasible Reasoning for Datalog - OpenUCT