Tutorial #1 – Cognitive Effectiveness of Representations for Process Mining
Jan Mendling, Djordje Djurica and Monika Malinova
Jan Mendling is a Full Professor with the Department of Computer Science at Humboldt-Universität zu Berlin, Germany
Monika Malinova is a postdoctoral researcher with the Institute for Information Business at the Vienna University of Economics and Business (WU), Austria.
Djordje Djurica is a Teaching and Research Associate with the Institute for Information Business at Vienna University of Economics and Business (WU Wien).
Abstract: Process mining produces a variety of outputs. Many of these outputs are visual diagrams as for example BPMN Diagrams, Petri nets, or Directly-Follows Graphs, to name just a few. The benefits of such diagrams are largely acclaimed in Business Process Management (BPM); however, research on this topic is fragmented. In this tutorial, we look at process diagrams from a holistic perspective. In the first part, we discuss the essential characteristics of diagrams generated by process mining techniques with a focus on purposes, properties, and the cognitive architecture framework CogniDia for processing them. In the second part, we present findings from a recent survey that we conducted on criteria for designing effective diagrams. These findings are mapped to the CogniDia framework and its cognitive processing steps of visual, verbal and semantic elements of process diagrams, as well as task processing. This tutorial is relevant for junior researchers and senior scholars. For junior researchers, it provides a compact overview of various cognitive theories that could be used to understand processing of process diagrams.
For senior scholars, it is interesting for its novel research insights of a survey into 220 guidelines for effective diagram processing, organized in the four large groups of effective visual processing, verbal processing, semantic processing, and task processing. For any researcher interested in process mining, the tutorial emphasizes the importance of assessing process mining outputs from the perspective of the analyst tasks at hand.
Tutorial #2 – Applications of Automated Planning for Business Process Management
Andrea Marrella and Tathagata Chakraborti
Andrea Marrella is Assistant Professor with Sapienza Università di Roma, Italy.
Tathagata Chakraborti works on planning and human-AI collaboration at IBM Research.
Abstract: Automated planning is the branch of Artificial Intelligence (AI) that concerns the synthesis of autonomous behaviours, consisting of strategies or action sequences (called plans) for specific classes of mathematical models represented in compact form. In recent years, the automated planning community has developed a plethora of planning systems known as planners, which embed very effective (i.e., scale up to large problems) search heuristics that have been employed to solve collections of challenging problems from many Computer Science domains. In this tutorial, we discuss how automated planning can be leveraged to enable new levels of automation and support for Business Process Management (BPM) in a theoretically grounded and domain-independent way. Specifically, we first describe how a researcher/practitioner should approach the task of encoding real-world problems as appropriate planning problems and under which conditions planning is feasible for solving them. Then, we discuss how instances of some well-known problems from the BPM literature can be represented as planning problems for which planners can find a correct solution in a finite amount of time. Finally, we show how to integrate the planning technology with traditional BPM systems.
Tutorial #3 – RuM: Declarative Process Mining, Distilled
Anti Alman, Claudio Di Ciccio, Fabrizio Maria Maggi, Marco Montali and Han van der Aa
Anti Alman is a Junior Research Fellow in Information Systems and a PhD student at the University of Tartu.
Claudio Di Ciccio is an Assistant Professor at the Department of Computer Science of the Sapienza University of Rome, Italy.
Fabrizio Maria Maggi is Associate Professor at the KRDB Research Centre for Knowledge and Data.
Marco Montali is Associate Professor and Vice-Dean of Teaching at the Faculty of Computer Science, Free University of Bozen-Bolzano, Italy.
Han van der Aa is a junior professor in the Data and Web Science Group at the University of Mannheim.
Abstract: Flexibility is a key characteristic of numerous business process management domains. In such settings, dynamic decisions can deeply affect process executions depending on the current circumstances. Therefore, the paths to fulfil the process goals are not fully predetermined. A common example is the adaptation of a standard treatment process to the needs of a specific patient. However, high flexibility does not mean chaos: a set of key process rules still delimits the execution space. For example, certain drugs have dangerous interactions and therefore cannot coexist in the same treatment case. One of the most renowned paradigms for the handling of flexibility is the declarative approach, which aims at defining processes through their core behavioural rules, thus leaving room for adaptation when necessary. To achieve its full potential, the declarative approach requires a paradigm shift in process thinking and, therefore, the support of novel tools. The goal of this tutorial is to distil the declarative approach down to its core concepts. We use the Declare language and the RuM toolkit as a concrete backbone. The first part of the tutorial focuses on the main principles and constructs of Declare. The second part provides practical examples and exercises using RuM.
Tutorial #4 – Artifact-driven Process Monitoring: a viable solution to continuously and autonomously monitor business processes
Giovanni Meroni ia a Postdoctoral Research Assistant at Politecnico di Milano.
Additional Material: https://tinyurl.com/smartifact
Abstract: Business process monitoring aims at identifying how well running processes are performing with respect to performance measures and objectives. By observing the execution of a process, process monitoring is also responsible for creating process traces, which can be subsequently used by process mining algorithms to gain further insights on the process.Among the various monitoring solutions, artifact-driven monitoring has been proposed as a viable solution to continuously and autonomously monitor business processes. By monitoring the changes in the physical and virtual objects (i.e., artifacts) participating in the process, artifact-driven monitoring can autonomously generate traces that include events related to semi-automatic and manual tasks. In this way, the operators responsible for such activities no longer have to send notifications to the monitoring platform. Also, by relying on a declarative representation of the process to monitor, artifact-driven monitoring can detecting violations in the execution flow as soon as they occur. In addition, artifact-driven monitoring can identify the process elements affected by a violation, and it can continue monitoring the process without human intervention.
This tutorial will firstly provide an introduction to process monitoring, and the recent advancements in this field. Then, the inner details on how artifact-driven monitoring works will be provided