Learning by Intervention in Simple Causal Domains

Katrine Bjørn Pedersen Thoft, Nina Gierasimczuk*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

We propose a framework for learning dependencies between variables in an environment with causal relations. We assume that the environment is fully observable and that the underlying causal structure is of a simple nature. We adapt the frameworks of the (epistemic) causal models from [4, 17], and propose a model inspired by action learning [6, 7]. We present two learning methods, using formal and algorithmic approaches. Our learning agents infer dependencies (atomic formulas of Dependence Logic) from observations of interventions on valuations (propositional states), and by doing so efficiently, they obtain insights into how to manipulate their surroundings to achieve goals.
Original languageEnglish
Title of host publicationProceedings of the 5th International Workshop of Dynamic Logic. New Trends and Applications : DaLí 2023
Volume14401
PublisherSpringer
Publication date2024
Pages104-118
ISBN (Print)978-3-031-51776-1
ISBN (Electronic)978-3-031-51777-8
DOIs
Publication statusPublished - 2024
EventDynamic Logic. New Trends and Applications 2023: DaLí - Tbilisi, Georgia
Duration: 15 Sept 202316 Sept 2023
Conference number: 5

Workshop

WorkshopDynamic Logic. New Trends and Applications 2023
Number5
Country/TerritoryGeorgia
CityTbilisi
Period15/09/202316/09/2023

Keywords

  • Causality
  • Casual models
  • Dependence models
  • Dependence logic
  • Learning by intervention
  • Finite identifiability
  • Artificial intelligence

Fingerprint

Dive into the research topics of 'Learning by Intervention in Simple Causal Domains'. Together they form a unique fingerprint.

Cite this