The ADRIA team proposed an original approach for modelling the causal relationships perceived by agents in response to a series of events, based on the  comparison of the agent’s beliefs about the normal course of events (represented using a non-monotonous consequence relationship) with the events that occurred. This modeling was validated on a psychological level, and compared with probabilistic modeling successfully.

Another approach involves the idea that verifying that an event can be considered a cause in a given scenario amounts to a scenario update problem in the event that the event has not occurred.

  • Jean-François Bonnefon, Rui Da Silva Neves, Didier Dubois, Henri Prade: Qualitative and quantitative conditions for the transitivity of perceived causation: – Theoretical and experimental results. Ann. Math. Artif. Intell. 64(2-3): 311-333, 2012.
  • Didier Dubois, Henri Prade. A glance at causality theories for Artificial Intelligence. In : A Guided Tour of Artificial Intelligence Research. Vol. 1 Knowledge Representation, Reasoning and Learning, (P. Marquis, O. Papini, H. Prade, eds.), Springer 2019.