Volodymyr Pavlyshyn·Oct 28Your Agent’s Memory Should Be Able to Say NoA user tells your assistant: Anna and Lena met in Berlin on day 2001, from 10:00 to 12:00. The model turns the sentence into a fact and…
InArtificial Intelligence in Plain EnglishbyVolodymyr Pavlyshyn·Oct 27Bulletproof Agent with Verifiable Certified OntologiesMost ontologies are described, not proved: OWL and SHACL tell you what a reasoner believes, but they do not hand you a certificate that you…
Volodymyr Pavlyshyn·May 26, 2025Emulating Dependent Types in Elpi: A Journey into Logic Programming and Type TheoryDependent types, where the type of a value can depend on the value itself, are a cornerstone of modern programming languages and proof…
InArtificial Intelligence in Plain EnglishbyVolodymyr Pavlyshyn·May 8, 2025Lambda Prolog and ELPI: Advancing Knowledge Graphs and Complex OntologiesIn a discussion of HOL and Dependent Types for ontologies, we start an interesting debate but miss some practical instruments.
