Case study providing
Formalisation of Context
In real-world
Applications
Context-based modelling of shared context of pair of actors in decision-making
Patrick Brézillon* and Patrick Humphreys**
*Sorbonne Université, Paris, France
**London School of Economics and Political Science, UK
Abstract
This paper presents a formalisation of context for use in real-world applications in coordination with concepts coming from Cognitive Sciences. Our AI approach is centred on the modelling of actor’s experience for accomplishing an activity in different contexts. An actor develops each time a mental model as an operational explanation of an activity in a specific context, and concretises experience in a mental representation containing all the mental models created in the different contexts where the activity was realised. The approach is based on activity modelling in a referential at four levels, namely conceptual, operational, implementation and environment levels. The Contextual-Graphs (CxG) formalism, which achieves the four components of the approach for one actor activity (CxG_1.0 version) and for group activity (CxG_2.0 version). The resulting software offers a new vision of context as proceduralised context leading to a definition of real time context and to a CxG-based simulation tool for following contextual reasoning during activity development, especially collaborative activity. The objective of the paper is the modelling of shared context between two actors for decision-making. The different aspects of the CxG formalism are discussed in the Anam Cara Ontology project that studies the interactions of nannies with their Anam Caras (soul friends).
Keywords: AI systems, decision-making, activity, mental model, contextual reasoning, contextual-graphs
