Causality and Hyperproperties

Norine Coenen

Hyperproperties specify relations between multiple execution traces of a system. This is necessary to express requirements like secrecy where for every possible value of a secret variable there has to exist a computation where the value is different while the observations made by an external observer are the same. Causal relations are often defined using counterfactuals. Intuitively, an action c is considered to be a cause for an effect e if there is one witness where c and e occur and another witness similar to the first where c and e both do not occur. Thus, causality itself is a hyperproperty that compares multiple system traces.

NII Shonan Meeting No. 139 on Causal Reasoning in Systems June 2019
Contact Data Privacy Policy Imprint
Home People Publications
More