Guiding the generation of counterfactual explanations through temporal background knowledge for Predictive Process Monitoring

arXiv:2403.11642v1 Announce Type: cross
Abstract: Counterfactual explanations suggest what should be different in the input instance to change the outcome of an AI system. When dealing with counterfactual explanations in the field of Predictive Process Monitoring, however, control flow relationships among events have to be carefully considered. A counterfactual, indeed, should not violate control flow relationships among activities (temporal background knowledege). Within the field of Explainability in Predictive Process Monitoring, there have been a series of works regarding counterfactual explanations …

abstract ai system arxiv change control counterfactual cs.ai cs.lg events flow however instance knowledge monitoring predictive process relationships temporal through type