Context-based modelling for nanny decision-making
In living sciences, the scientific approach for developing knowledge aims to formalise concepts by confrontation of a model to observable phenomena of a real system (Brézillon 1983). It is called a hypothetical-deductive approach with an inductive complement when the model fails to explain observable phenomena: the ideas need to be revised and the model improved. We adapt this scientific approach for modelling in AI how an actor carries out an activity in a given situation with a local environment and the needed resources for the realization of the activity (Brézillon, to appear). Such conditions constitute the context of the activity.
Our approach is based on a modelling of actor’s activity at four levels, namely, conceptual, operational, implementation and environment levels. We also consider activity modelling carried out by a group of actors by extending the framework used for one actor activity. A context-based formalism of representation plays an important role of “concept revealer” in an activity model at different levels from concepts to an operational model and finally to an implementation to be fed at to context sources in activity environment. Reasoning is defined in various disciplines (more or less formal) along different priorities: formal exploration of logical rules, psychological mechanisms, or even practical applications. Reasoning consists of starting from collecting, assembling and structuring contextual knowledge and information for making a valid decision that maximises the means to use for a given goal. Thus, decision-making being central in activities.
Section 2 presents our approach for modelling actor activity at four levels, from the more abstract to the more concrete one, starting from concepts used in our modelling with refinement of the concepts at the lower levels. Section 3 presents group-activity modelling as an extension of the actor-activity infrastructure. Section 4 illustrates the possibilities of the CxG formalism on the example of « nanny-anam cara interactions ». Section 5 presents close works related to our approach, and Section 6 concludes by showing an opening for other research on this topic and for merging the CxG formalism with new promising ways.
Our approach is based on a modelling of actor’s activity at four levels, namely, conceptual, operational, implementation and environment levels. We also consider activity modelling carried out by a group of actors by extending the framework used for one actor activity. A context-based formalism of representation plays an important role of “concept revealer” in an activity model at different levels from concepts to an operational model and finally to an implementation to be fed at to context sources in activity environment. Reasoning is defined in various disciplines (more or less formal) along different priorities: formal exploration of logical rules, psychological mechanisms, or even practical applications. Reasoning consists of starting from collecting, assembling and structuring contextual knowledge and information for making a valid decision that maximises the means to use for a given goal. Thus, decision-making being central in activities.
Section 2 presents our approach for modelling actor activity at four levels, from the more abstract to the more concrete one, starting from concepts used in our modelling with refinement of the concepts at the lower levels. Section 3 presents group-activity modelling as an extension of the actor-activity infrastructure. Section 4 illustrates the possibilities of the CxG formalism on the example of « nanny-anam cara interactions ». Section 5 presents close works related to our approach, and Section 6 concludes by showing an opening for other research on this topic and for merging the CxG formalism with new promising ways.
You can explore the key features of Context-based modelling for nanny decision-making by clicking on any of the section titles Identified in yellow
Modelling levels for actor activity
Extension of the CxG formalism to group activity
Nanny-Anam Cara interactions example Part one
Nanny-Anam Cara interactions example Part Two
Nanny-Anam Cara interactions example: Conclusion
Extension of the CxG formalism to group activity
Nanny-Anam Cara interactions example Part one
Nanny-Anam Cara interactions example Part Two
Nanny-Anam Cara interactions example: Conclusion
