Predictive Process Monitoring (PPM) forecasts future events of ongoing business processes based on event logs. However, deep learning PPM approaches are often limited by labeled data scarcity and the low variability of real-world logs. To address this, we introduce SiamSA-PPM, a self-supervised framework combining Siamese learning with statistically grounded data augmentation. It employs novel transformation methods that leverage control-flow semantics and frequent behavioral patterns to generate realistic, semantically valid trace variants, learning generalizable prefix representations without labeled supervision. Extensive experiments on real-world event logs with a wide range of variability, incorporating exhaustive hyperparameter tuning, demonstrate the value of this process-aware augmentation. Compared to state-of-the-art baselines, SiamSA-PPM achieves competitive or superior predictive performance in both next activity and final outcome prediction tasks. An open source implementation of our method is publicly available under https://github.com/SvStraten/SiamSA-PPM-Extension..
Automated paraphrasing has numerous applications such as query-answering, dialog systems, and machine translation. Recently, due to the limitations of automatic paraphrasing systems, crowdsourcing emerged as an alternative for obtaining corpora of paraphrases. Since crowdsourced paraphrases are generated by crowd workers with varied skills and motivations, such paraphrases may suffer from various quality issues (e.g. grammatical and spelling errors). Such quality issues stem from the lack of quality assessment methods for open-ended text in crowdsourcing platforms. Thus unqualified or malicious workers may generate erroneous paraphrases. Crowdsourced paraphrases must be assessed to filter low quality paraphrases, imposing an extra cost. Detecting such erroneous paraphrases requires an understanding of how malicious workers generate paraphrases. In this paper, we propose an effective approach to detect malicious worker behaviors based on a comprehensive analysis of cheating behavior characteristics. We investigate what models and features are effective in detecting such behaviors, and explore mechanisms and crowdsourcing prompts that could lead to an effective integration in an actual crowdsourcing task. Our results point to a significant reduction of cheating behavior (from 18% to 4.4%) in applying automatic malicious worker behavior detection, and highlight the impact of the detection (and detection mistakes) on workers corrective measures.
Process mining enables organizations to discover, monitor, and analyze their work processes based on data. A fundamental requirement for initiating a process mining project is the availability of an event log, which is not always readily available. In such cases, extracting an event log typically involves various time-consuming tasks, such as writing custom structured query language (SQL) scripts to extract relevant data into an event log format from a relational database. In this work, we explore the potential of large language models (LLMs) to support event log extraction for process mining by leveraging LLMs' ability to produce SQL scripts. We evaluate the effectiveness of LLMs in assisting this process and analyze their performance across a range of scenarios. Despite the inherent non-determinism of LLM outputs, our findings highlight the potential of future LLM-assisted tools in automating event log extraction, particularly when provided with the appropriate domain and data knowledge context. The implementation of such tools could democratize access to process mining by reducing the need for specialized technical expertise for producing relational database query scripts and minimizing manual effort.
Effective business process execution requires the integration of both process logic and business data. While recent approaches explore the potential of Large Language Models (LLMs) in automating process modeling, their applicability is limited in real-world scenarios where textual descriptions - often authored by non-experts - are complex or incomplete. Moreover, these works primarily focus on the control-flow perspective and overlook the critical role of data modeling and execution. In this paper, we propose a hybrid and decomposed approach to automatically generate executable process and data models from text using LLMs. Our method modularizes the task: the LLM clarifies and enriches the description, then extracts both process and data elements, which are combined into a unified model. Structured algorithms ensure robust and executable outputs. Evaluation results demonstrate that our approach improves model completeness, clarity, and efficiency compared to existing methods.
Predictive process monitoring (PPM) focuses on predicting future process trajectories, including next activity predictions. This is crucial in dynamic environments where processes change or face uncertainty. However, current frameworks often assume a static environment, overlooking dynamic characteristics and concept drifts. This results in catastrophic forgetting, where training while focusing merely on new data distribution negatively impacts the performance on previously learned data distributions. Continual learning addresses, among others, the challenges related to mitigating catastrophic forgetting. This paper proposes a novel approach called Continual Next Activity Prediction with Prompts (CNAPwP), which adapts the DualPrompt algorithm for next activity prediction to improve accuracy and adaptability while mitigating catastrophic forgetting. We introduce new datasets with recurring concept drifts, alongside a task-specific forgetting metric that measures the prediction accuracy gap between initial occurrence and subsequent task occurrences. Extensive testing on three synthetic and two real-world datasets representing several setups of recurrent drifts shows that CNAPwP achieves SOTA or competitive results compared to five baselines, demonstrating its potential applicability in real-world scenarios. An open-source implementation of our method, together with the datasets and results, is available at: https://github.com/SvStraten/CNAPwP.
