Causal learning in a quantum world

Date(s) : 14/09/2018
14 h 00 min - 15 h 00 min

Causal reasoning is essential to how we understand and model the world. Quite recently, a general theory of causal modelling has been developed, with the ambitious goal to provide a unified framework for causal discovery across all disciplines. The practical applications of this framework are still unclear, as building algorithms that learn causal structures, as opposed to sheer correlations, remains challenging. However, where the framework fails most dramatically is quantum physics, where causal models are unable to reproduce simple scenarios. What is worse, the ordinary quantum formalism does not support causal inference: one cannot meaningfully ask, given two events, which is the cause and which the effect. This is particularly troubling, as future quantum technologies might require systematic methods to discover cause-effect relations in large networks. I will present a novel formalism that makes quantum inference a well-defined task, with possible applications to the analysis of complex processes. I will discuss how machine-learning techniques can help solving this task in the absence of complete information.

Catégories Pas de Catégories

Retour en haut 

Secured By miniOrange