Search

The field of predictive process monitoring focuses on case-level models to predict a single specific outcome such as a particular objective, (remaining) time, or next activity/remaining sequence. Recently, a longer-horizon, model-wide approach has been proposed in the form of process model forecasting, which predicts the future state of a whole process model through the forecasting of all activity-to-activity relations at once using time series forecasting. This paper introduces the concept of predictive process model monitoring which sits in the middle of both predictive process monitoring and process model forecasting. Concretely, by modelling a process model as a set of constraints being present between activities over time, we can capture more detailed information between activities compared to process model forecasting, while being compatible with typical predictive process monitoring objectives which are often expressed in the same language as these constraints. To achieve this, Processes-As-Movies (PAM) is introduced, i.e., a novel technique capable of jointly mining and predicting declarative process constraints between activities in various windows of a process' execution. PAM predicts what declarative rules hold for a trace (objective-based), which also supports the prediction of all constraints together as a process model (model-based). Various recurrent neural network topologies inspired by video analysis tailored to temporal high-dimensional input are used to model the process model evolution with windows as time steps, including encoder–decoder long short-term memory networks, and convolutional long short-term memory networks. Results obtained over real-life event logs show that these topologies are effective in terms of predictive accuracy and precision. [ABSTRACT FROM AUTHOR]

Goal Recognition (GR) is a research problem that studies ways to infer the goal of an intelligent agent based on its observed behavior and knowledge of the environment in which the agent operates. A common assumption of GR is that the environment is static. However, in many real-world scenarios, for example, recognizing customers' preferences, it is necessary to recognize the goals of multiple agents or multiple goals of a single agent over an extended period. Therefore, it is reasonable to expect the environment to change throughout a series of goal recognition tasks. This paper presents three process mining-based solutions to the problem of adaptive GR in a changing environment implemented as different control strategies of a system for solving standard GR problems. As a standard GR system that gets controlled, we use the system grounded in process mining techniques, as it can adjust its internal GR mechanisms based on data collected while observing the operating agents. We evaluated our control strategies over synthetic and real-world datasets. The synthetic datasets were generated using the extended version of the Goal Recognition Amidst Changing Environments (GRACE) tool. The datasets account for different types of changes and drifts in the environment. The evaluation results demonstrate a trade-off between the GR performance over time and the effort invested in adaptations of the GR mechanisms of the system, showing that few well-planned adaptations can lead to a consistently high GR performance. [ABSTRACT FROM AUTHOR]

The field of predictive process monitoring focuses on case-level models to predict a single specific outcome such as a particular objective, (remaining) time, or next activity/remaining sequence. Recently, a longer-horizon, model-wide approach has been proposed in the form of process model forecasting, which predicts the future state of a whole process model through the forecasting of all activity-to-activity relations at once using time series forecasting. This paper introduces the concept of predictive process model monitoring which sits in the middle of both predictive process monitoring and process model forecasting. Concretely, by modelling a process model as a set of constraints being present between activities over time, we can capture more detailed information between activities compared to process model forecasting, while being compatible with typical predictive process monitoring objectives which are often expressed in the same language as these constraints. To achieve this, Processes-As-Movies (PAM) is introduced, i.e., a novel technique capable of jointly mining and predicting declarative process constraints between activities in various windows of a process' execution. PAM predicts what declarative rules hold for a trace (objective-based), which also supports the prediction of all constraints together as a process model (model-based). Various recurrent neural network topologies inspired by video analysis tailored to temporal high-dimensional input are used to model the process model evolution with windows as time steps, including encoder–decoder long short-term memory networks, and convolutional long short-term memory networks. Results obtained over real-life event logs show that these topologies are effective in terms of predictive accuracy and precision. [ABSTRACT FROM AUTHOR]

A key challenge in data mining, machine learning and artificial intelligence concerns data quality. Process mining is not an exception. A range of data quality problems exists in process data, some of which caused by activity labels. While automated techniques can be used for remedying data quality problems, they are effective up to a point and the aid of domain experts is required (cf. the "human-in-the-loop" paradigm in artificial intelligence). They are eminently suited to fix incorrect labels, but hard to engage as the repair task can be time-consuming and tedious. Gamification may offer a promising solution to this challenge. In this paper, we examine what motivational drives can be exploited to gamify activity-label repair, specifically those with identical semantics but different syntax. We conducted two experiments. First we recruited experts from the insurance domain to repair labels using our gamified system. Results show that log quality was improved, and participants had a positive experience interacting with the system. To validate the generalisability of the approach, our second experiment involved 30 participants who repaired a real-life medical log. We found promising results on quality improvement of the log and comparison with automatic approaches. This work contributes to process mining by improving activity-label quality in event logs and by turning data cleaning from the least enjoyable task to a fun experience for users. By bringing together elements from gamification, crowdsourcing ("humans-in-the-loop"), and process mining, this study contributes to improving the reliability of data-driven decision making in machine learning and artificial intelligence. • We investigated the impact of gamification on expert engagement in label repair. • We examined experts' motivations to contribute knowledge. • Two experiments were conducted in the insurance and medical domains. • We observed improvement in the quality of the event log. • Epic meaning, development, and relatedness along with other drives motivate experts. [ABSTRACT FROM AUTHOR]

Process mining is an interdisciplinary field that combines Artificial Intelligence and Business Process Management to extract insights from historical event data. Knowledge-intensive processes, which predominantly involve knowledge work, are often inadequately monitored by process-aware information systems. Consequently, the event data necessary for applying process mining techniques are frequently unavailable. Emails are widely used in knowledge-intensive processes for scheduling meetings, sharing documents, and reporting on the completion of outstanding tasks, which makes them suitable candidates for replacing event logs as the primary data source. In this work, we focus on the task of extracting the set of next recommended activities from incoming emails. Yet, we face two major challenges. Firstly, emails do not express process information explicitly but rather contain subtext that implies what the next best actions would be. Secondly, email data lacks domain-specific labels that would enable the use of machine learning. We overcome these limitations by utilizing an email taxonomy to represent user intents, thus bridging the gap between textual information and process semantics, as well as leveraging pre-trained transformer models applied in zero-shot and few-shot settings that require little to no labeled email data. An evaluation of our method on real-world unlabeled email communications demonstrates its effectiveness in recognizing intents and extracting activities. [ABSTRACT FROM AUTHOR]