arXiv subjects

Daniel Schuster

Publications and source records attributed to Daniel Schuster.

At least 19 recordsLinked to original sources

arXiv2026

Automated Event Log Generation from Unstructured Text Using Finetuned LLMs

Process mining (PM) provides a powerful framework for discovering and optimizing operational processes from event data. However, the efficacy of PM techniques is strictly predicated on the availability of structured event logs. Thus far, event logs have often been laboriously created by domain and process mining experts. This costly effort causes large portions of organizational knowledge, including incident tickets, manuals, and textual reports, to remain underutilized. We address this bottleneck by investigating the efficacy of Large Language Models (LLMs) as automated data translators. We propose a scalable framework that leverages LLMs as data translators to bridge the gap between unstructured textual resources and structured event data. We finetune LLMs on a newly created text-to-log dataset, demonstrating that the resulting models can extract high-fidelity event logs from unstructured resources. Our results show that this finetuning approach outperforms few-shot or zero-shot prompting by a large amount, highlighting finetuning as a necessary pre-condition for generating reliable event data. We conclude that our method provides a promising pipeline for making previously unused data available to process mining ecosystems, effectively expanding the possibilities of using PM to further investigate organizational workflows.

cs.AI↗

arXiv2025

Computing Alignments for Partially-ordered Traces Through Petri Net Unfoldings

Conformance checking techniques aim to provide diagnostics on the conformity between process models and event data. Conventional methods, such as trace alignments, assume strict total ordering of events, leading to inaccuracies when timestamps are overlapping, coarse, or missing. In contrast, existing methods that support partially ordered events rely upon the interleaving semantics of Petri nets, the reachability graphs, which suffer from the state space explosion problem. %Besides, this view also forces the solution to create a partially ordered alignment almost artificially. This paper proposes an improved approach to conformance checking based upon partially ordered event data by utilizing Petri net unfolding, which leverages partial-order semantics of Petri nets to represent concurrency and uncertainty in event logs more effectively. Unlike existing methods, our approach offers a streamlined one-step solution, improving efficiency in the computation of alignments. Additionally, we introduce a novel visualization technique for partially ordered unfolding-based alignments. We implement unfolding-based alignments with its user-friendly insights in a conformance analysis tool. Our experimental evaluation, conducted on synthetic and real-world event logs, demonstrates that the unfolding-based approach is particularly robust in handling high degrees of parallelism and complexity in process models.

cs.DB↗

arXiv2024

Evaluating Large Language Models on Business Process Modeling: Framework, Benchmark, and Self-Improvement Analysis

Large Language Models (LLMs) are rapidly transforming various fields, and their potential in Business Process Management (BPM) is substantial. This paper assesses the capabilities of LLMs on business process modeling using a framework for automating this task, a comprehensive benchmark, and an analysis of LLM self-improvement strategies. We present a comprehensive evaluation of 16 state-of-the-art LLMs from major AI vendors using a custom-designed benchmark of 20 diverse business processes. Our analysis highlights significant performance variations across LLMs and reveals a positive correlation between efficient error handling and the quality of generated models. It also shows consistent performance trends within similar LLM groups. Furthermore, we investigate LLM self-improvement techniques, encompassing self-evaluation, input optimization, and output optimization. Our findings indicate that output optimization, in particular, offers promising potential for enhancing quality, especially in models with initially lower performance. Our contributions provide insights for leveraging LLMs in BPM, paving the way for more advanced and automated process modeling techniques.

cs.DB↗

arXiv2024

Leveraging Large Language Models for Enhanced Process Model Comprehension

In Business Process Management (BPM), effectively comprehending process models is crucial yet poses significant challenges, particularly as organizations scale and processes become more complex. This paper introduces a novel framework utilizing the advanced capabilities of Large Language Models (LLMs) to enhance the interpretability of complex process models. We present different methods for abstracting business process models into a format accessible to LLMs, and we implement advanced prompting strategies specifically designed to optimize LLM performance within our framework. Additionally, we present a tool, AIPA, that implements our proposed framework and allows for conversational process querying. We evaluate our framework and tool by i) an automatic evaluation comparing different LLMs, model abstractions, and prompting strategies and ii) a user study designed to assess AIPA's effectiveness comprehensively. Results demonstrate our framework's ability to improve the accessibility and interpretability of process models, pioneering new pathways for integrating AI technologies into the BPM field.

cs.DB↗

arXiv2024

Accelerating Aeroelastic UVLM Simulations by Inexact Newton Algorithms

We consider the aeroelastic simulation of flexible mechanical structures submerged in subsonic fluid flows at low Mach numbers. The nonlinear kinematics of flexible bodies are described in the total Lagrangian formulation and discretized by finite elements. The aerodynamic loads are computed using the unsteady vortex-lattice method wherein a free wake is tracked over time. Each implicit time step in the dynamic simulation then requires solving a nonlinear equation system in the structural variables with additional aerodynamic load terms. Our focus here is on the efficient numerical solution of this system by accelerating the Newton algorithm. The particular structure of the aeroelastic nonlinear system suggests the structural derivative as an approximation to the full derivative in the linear Newton system. We investigate and compare two promising algorithms based in this approximation, a quasi-Newton type algorithm and a novel inexact Newton algorithm. Numerical experiments are performed on a flexible plate and on a wind turbine. Our computational results show that the approximation can indeed accelerate the Newton algorithm substantially. Surprisingly, the theoretically preferable inexact Newton algorithm is much slower than the quasi-Newton algorithm, which motivates further research to speed up derivative evaluations.

math.NA↗

arXiv2024

Process Modeling With Large Language Models

In the realm of Business Process Management (BPM), process modeling plays a crucial role in translating complex process dynamics into comprehensible visual representations, facilitating the understanding, analysis, improvement, and automation of organizational processes. Traditional process modeling methods often require extensive expertise and can be time-consuming. This paper explores the integration of Large Language Models (LLMs) into process modeling to enhance the accessibility of process modeling, offering a more intuitive entry point for non-experts while augmenting the efficiency of experts. We propose a framework that leverages LLMs for the automated generation and iterative refinement of process models starting from textual descriptions. Our framework involves innovative prompting strategies for effective LLM utilization, along with a secure model generation protocol and an error-handling mechanism. Moreover, we instantiate a concrete system extending our framework. This system provides robust quality guarantees on the models generated and supports exporting them in standard modeling notations, such as the Business Process Modeling Notation (BPMN) and Petri nets. Preliminary results demonstrate the framework's ability to streamline process modeling tasks, underscoring the transformative potential of generative AI in the BPM field.

cs.SE↗

arXiv2024

Evaluating Large Language Models in Process Mining: Capabilities, Benchmarks, and Evaluation Strategies

Using Large Language Models (LLMs) for Process Mining (PM) tasks is becoming increasingly essential, and initial approaches yield promising results. However, little attention has been given to developing strategies for evaluating and benchmarking the utility of incorporating LLMs into PM tasks. This paper reviews the current implementations of LLMs in PM and reflects on three different questions. 1) What is the minimal set of capabilities required for PM on LLMs? 2) Which benchmark strategies help choose optimal LLMs for PM? 3) How do we evaluate the output of LLMs on specific PM tasks? The answer to these questions is fundamental to the development of comprehensive process mining benchmarks on LLMs covering different tasks and implementation paradigms.

cs.DB↗

arXiv2024

ProMoAI: Process Modeling with Generative AI

ProMoAI is a novel tool that leverages Large Language Models (LLMs) to automatically generate process models from textual descriptions, incorporating advanced prompt engineering, error handling, and code generation techniques. Beyond automating the generation of complex process models, ProMoAI also supports process model optimization. Users can interact with the tool by providing feedback on the generated model, which is then used for refining the process model. ProMoAI utilizes the capabilities LLMs to offer a novel, AI-driven approach to process modeling, significantly reducing the barrier to entry for users without deep technical knowledge in process modeling.

cs.DB↗

arXiv2023

Abstractions, Scenarios, and Prompt Definitions for Process Mining with LLMs: A Case Study

Large Language Models (LLMs) are capable of answering questions in natural language for various purposes. With recent advancements (such as GPT-4), LLMs perform at a level comparable to humans for many proficient tasks. The analysis of business processes could benefit from a natural process querying language and using the domain knowledge on which LLMs have been trained. However, it is impossible to provide a complete database or event log as an input prompt due to size constraints. In this paper, we apply LLMs in the context of process mining by i) abstracting the information of standard process mining artifacts and ii) describing the prompting strategies. We implement the proposed abstraction techniques into pm4py, an open-source process mining library. We present a case study using available event logs. Starting from different abstractions and analysis questions, we formulate prompts and evaluate the quality of the answers.

cs.DB↗

arXiv2022

Control-Flow-Based Querying of Process Executions from Partially Ordered Event Data

Event logs, as viewed in process mining, contain event data describing the execution of operational processes. Most process mining techniques take an event log as input and generate insights about the underlying process by analyzing the data provided. Consequently, handling large volumes of event data is essential to apply process mining successfully. Traditionally, individual process executions are considered sequentially ordered process activities. However, process executions are increasingly viewed as partially ordered activities to more accurately reflect process behavior observed in reality, such as simultaneous execution of activities. Process executions comprising partially ordered activities may contain more complex activity patterns than sequence-based process executions. This paper presents a novel query language to call up process executions from event logs containing partially ordered activities. The query language allows users to specify complex ordering relations over activities, i.e., control flow constraints. Evaluating a query for a given log returns process executions satisfying the specified constraints. We demonstrate the implementation of the query language in a process mining tool and evaluate its performance on real-life event logs.

cs.DB↗

arXiv2022

A Framework for Extracting and Encoding Features from Object-Centric Event Data

Traditional process mining techniques take event data as input where each event is associated with exactly one object. An object represents the instantiation of a process. Object-centric event data contain events associated with multiple objects expressing the interaction of multiple processes. As traditional process mining techniques assume events associated with exactly one object, these techniques cannot be applied to object-centric event data. To use traditional process mining techniques, the object-centric event data are flattened by removing all object references but one. The flattening process is lossy, leading to inaccurate features extracted from flattened data. Furthermore, the graph-like structure of object-centric event data is lost when flattening. In this paper, we introduce a general framework for extracting and encoding features from object-centric event data. We calculate features natively on the object-centric event data, leading to accurate measures. Furthermore, we provide three encodings for these features: tabular, sequential, and graph-based. While tabular and sequential encodings have been heavily used in process mining, the graph-based encoding is a new technique preserving the structure of the object-centric event data. We provide six use cases: a visualization and a prediction use case for each of the three encodings. We use explainable AI in the prediction use cases to show the utility of both the object-centric features and the structure of the sequential and graph-based encoding for a predictive model.

cs.LG↗

arXiv2022

Conformance Checking for Trace Fragments Using Infix and Postfix Alignments

Conformance checking deals with collating modeled process behavior with observed process behavior recorded in event data. Alignments are a state-of-the-art technique to detect, localize, and quantify deviations in process executions, i.e., traces, compared to reference process models. Alignments, however, assume complete process executions covering the entire process from start to finish or prefixes of process executions. This paper defines infix/postfix alignments, proposes approaches to their computation, and evaluates them using real-life event data.

cs.SE↗

arXiv2022

Defining Cases and Variants for Object-Centric Event Data

The execution of processes leaves traces of event data in information systems. These event data can be analyzed through process mining techniques. For traditional process mining techniques, one has to associate each event with exactly one object, e.g., the company's customer. Events related to one object form an event sequence called a case. A case describes an end-to-end run through a process. The cases contained in event data can be used to discover a process model, detect frequent bottlenecks, or learn predictive models. However, events encountered in real-life information systems, e.g., ERP systems, can often be associated with multiple objects. The traditional sequential case concept falls short of these object-centric event data as these data exhibit a graph structure. One might force object-centric event data into the traditional case concept by flattening it. However, flattening manipulates the data and removes information. Therefore, a concept analogous to the case concept of traditional event logs is necessary to enable the application of different process mining tasks on object-centric event data. In this paper, we introduce the case concept for object-centric process mining: process executions. These are graph-based generalizations of cases as considered in traditional process mining. Furthermore, we provide techniques to extract process executions. Based on these executions, we determine equivalent process behavior with respect to an attribute using graph isomorphism. Equivalent process executions with respect to the event's activity are object-centric variants, i.e., a generalization of variants in traditional process mining. We provide a visualization technique for object-centric variants. The contribution's scalability and efficiency are extensively evaluated. Furthermore, we provide a case study showing the most frequent object-centric variants of a real-life event log.

cs.DB↗

arXiv2021

Visualizing Trace Variants From Partially Ordered Event Data

Executing operational processes generates event data, which contain information on the executed process activities. Process mining techniques allow to systematically analyze event data to gain insights that are then used to optimize processes. Visual analytics for event data are essential for the application of process mining. Visualizing unique process executions -- also called trace variants, i.e., unique sequences of executed process activities -- is a common technique implemented in many scientific and industrial process mining applications. Most existing visualizations assume a total order on the executed process activities, i.e., these techniques assume that process activities are atomic and were executed at a specific point in time. In reality, however, the executions of activities are not atomic. Multiple timestamps are recorded for an executed process activity, e.g., a start-timestamp and a complete-timestamp. Therefore, the execution of process activities may overlap and, thus, cannot be represented as a total order if more than one timestamp is to be considered. In this paper, we present a visualization approach for trace variants that incorporates start- and complete-timestamps of activities.

cs.DB↗

arXiv2021

Freezing Sub-Models During Incremental Process Discovery: Extended Version

Process discovery aims to learn a process model from observed process behavior. From a user's perspective, most discovery algorithms work like a black box. Besides parameter tuning, there is no interaction between the user and the algorithm. Interactive process discovery allows the user to exploit domain knowledge and to guide the discovery process. Previously, an incremental discovery approach has been introduced where a model, considered to be under construction, gets incrementally extended by user-selected process behavior. This paper introduces a novel approach that additionally allows the user to freeze model parts within the model under construction. Frozen sub-models are not altered by the incremental approach when new behavior is added to the model. The user can thus steer the discovery algorithm. Our experiments show that freezing sub-models can lead to higher quality models.

cs.LG↗

arXiv2021

Cortado---An Interactive Tool for Data-Driven Process Discovery and Modeling

Process mining aims to diagnose and improve operational processes. Process mining techniques allow analyzing the event data generated and recorded during the execution of (business) processes to gain valuable insights. Process discovery is a key discipline in process mining that comprises the discovery of process models on the basis of the recorded event data. Most process discovery algorithms work in a fully automated fashion. Apart from adjusting their configuration parameters, conventional process discovery algorithms offer limited to no user interaction, i.e., we either edit the discovered process model by hand or change the algorithm's input by, for instance, filtering the event data. However, recent work indicates that the integration of domain knowledge in (semi-)automated process discovery algorithms often enhances the quality of the process models discovered. Therefore, this paper introduces Cortado, a novel process discovery tool that leverages domain knowledge while incrementally discovering a process model from given event data. Starting from an initial process model, Cortado enables the user to incrementally add new process behavior to the process model under construction in a visual and intuitive manner. As such, Cortado unifies the world of manual process modeling with that of automated process discovery.

cs.DB↗

arXiv2020

Scalable Online Conformance Checking Using Incremental Prefix-Alignment Computation

Conformance checking techniques aim to collate observed process behavior with normative/modeled process models. The majority of existing approaches focuses on completed process executions, i.e., offline conformance checking. Recently, novel approaches have been designed to monitor ongoing processes, i.e., online conformance checking. Such techniques detect deviations of an ongoing process execution from a normative process model at the moment they occur. Thereby, countermeasures can be taken immediately to prevent a process deviation from causing further, undesired consequences. Most online approaches only allow to detect approximations of deviations. This causes the problem of falsely detected deviations, i.e., detected deviations that are actually no deviations. We have, therefore, recently introduced a novel approach to compute exact conformance checking results in an online environment. In this paper, we focus on the practical application and present a scalable, distributed implementation of the proposed online conformance checking approach. Moreover, we present two extensions to said approach to reduce its computational effort and its practical applicability. We evaluate our implementation using data sets capturing the execution of real processes.

cs.LO↗

arXiv2020

Alignment Approximation for Process Trees

Comparing observed behavior (event data generated during process executions) with modeled behavior (process models), is an essential step in process mining analyses. Alignments are the de-facto standard technique for calculating conformance checking statistics. However, the calculation of alignments is computationally complex since a shortest path problem must be solved on a state space which grows non-linearly with the size of the model and the observed behavior, leading to the well-known state space explosion problem. In this paper, we present a novel framework to approximate alignments on process trees by exploiting their hierarchical structure. Process trees are an important process model formalism used by state-of-the-art process mining techniques such as the inductive mining approaches. Our approach exploits structural properties of a given process tree and splits the alignment computation problem into smaller sub-problems. Finally, sub-results are composed to obtain an alignment. Our experiments show that our approach provides a good balance between accuracy and computation time.

cs.DB↗