Argumentation meets Process Mining: an architecture for log interpretation Bettina Fazzinga 1 , Sergio Flesca 2 , Filippo Furfaro 2 and Luigi Pontieri 3 1 DICES - University of Calabria, Italy 2 DIMES - University of Calabria, Italy 3 ICAR - CNR, Italy Abstract We consider the scenario where there is an abstraction gap between the “low-level" events composing the traces in a business process log and the “high-level" activities in terms of which the analysts typically reason on the process behavior. We address the online interpretation problem of translating the event that has just been generated within a business process into the step of the activity instance it corresponds to. We present the architecture of a novel tool that models this interpretation problem as a dispute, encoded into an Abstract Argumentation Framework (AAF) [1], and that translates the computation of the valid interpretations, as well of the explanations on why the other interpretations are not valid, into instances of the AAF acceptance problem. 1. Introduction Thanks to the increasing difusion of automated tracing systems, the analysis of log data describing executions of business processes has gained momentum with the growth of the Process Mining research feld [2]. However, all the approaches and tools developed in this feld require that each log event can be mapped to well-defned activities, corresponding to some high-level view of the process. As a matter of fact, this assumption often does not hold in practice: in the logs of many processes, the events just represent low-level operations, with no clear reference to the business activities that were carried out through these operations, as shown in the following example. Example Consider the scenario of a hospital where patient medical records are stored by keeping track of the low-level events describing the exams and the checks performed by doctors and nurses. Suppose that a trace consists of Φ= e 1 ,e 2 ,e 3 ,e 4 , where e 1 is the event Blood sample taken, e 2 is Blood pressure measurement, e 3 is Temperature measurement, and e 4 is Cannula insertion, and that each of the frst 3 events can be performed during any of the high- level activities A 1 =pre-hospitalization, A 2 =pre-surgery, A 3 =post-surgery, while e 4 can be performed only during activity A 2 . In order to reconstruct the medical history of patients, there 5th Workshop on Advances in Argumentation in Artifcial Intelligence (AI 3 2021)  bettina.fazzinga@unical.it (B. Fazzinga); fesca@dimes.unical.it (S. Flesca); furfaro@dimes.unical.it (F. Furfaro); pontieri@icar.cnr.it (L. Pontieri)  0000-0001-8611-2377 (B. Fazzinga); 0000-0002-4164-940X (S. Flesca); 0000-0001-5145-1301 (F. Furfaro); 0000-0003-4513-0362 (L. Pontieri) © 2021 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings http://ceur-ws.org ISSN1613-0073 CEUR Workshop Proceedings (CEUR-WS.org)