Please contact me per email at least one day before your intended appointment.
Working Area: ICT Platform for Manufacturing
Production companies generate a wealth of data in various areas of the manufacturing processes, which can provide them with valuable insights for optimizing their processes, plants or systems. The effective exploitation of this industrial data requires the use of analytical techniques, including machine learning (ML) algorithms. Various components of an analysis solution must be selected and configured, such as data preparation techniques, IT computing infrastructure and the ML algorithms themselves. A selection of these components based on purely technical information can impair the performance of the resulting analysis solution, since the domain-specific objectives and requirements of the use case have not been explicitly considered.
In this PhD project, domain-specific factors for the selection and configuration of suitable ML algorithms in production are investigated. A framework proposed on the basis of this investigation allows the specification of all components required for an analysis solution. In addition to the data, the IT computing infrastructure and the ML-algorithms, suitable specifications can also be used to define the problem of the analysis solution via domain-specific objectives and requirements. The framework also allows to involve different experts from the respective application domain and from computer science or data sciences in the development process and to structure their cooperation better. Thus, cross-references between the specifications of the problem, data, IT computing infrastructure and ML algorithms can be defined in order to design, implement and validate analysis solutions in a holistic way.
- Villanueva Zacarias, A. G., Weber, C., Reimann, P., Mitschang, B. (2021). AssistML: A concept to recommend ML solutions for predictive use cases. 2021 IEEE 8th International Conference on Data Science and Advanced Analytics (DSAA), 2021, pp. 1-12, DOI: 10.1109/DSAA53316.2021.9564168.
- Villanueva Zacarias, A. G., Ghabri, R., Reimann, P. (2021). Leveraging Axiomatic Design to Improve the Design Process of Machine Learning Solutions in Manufacturing. Accepted for publication in a special issue of the International Journal of Semantic Computing (IJSC). To appear.
- Villanueva Zacarias, A. G., Ghabri, R. & Reimann, P. (2020). AD4ML: Axiomatic Design to Specify Machine Learning Solutions for Manufacturing. 2020 IEEE 21st International Conference on Information Reuse and Integration for Data Science (IRI) (p./pp. 148-155), August, 2020. DOI: 10.1109/IRI49571.2020.00029
- Villanueva Zacarias, A. G., Reimann, P. & Mitschang, B. (2018). A framework to guide the selection and configuration of machine-learning-based data analytics solutions in manufacturing. Procedia CIRP, 72, 153--158. DOI: 10.1016/j.procir.2018.03.215
- Villanueva Zacarias, A. G., Kassner, L. & Mitschang, B. (2017). Exploring Text Classification Configurations - A Bottom-up Approach to Customize Text Classifiers based on the Visualization of Performance. Proceedings of the 19th International Conference on Enterprise Information Systems, : SCITEPRESS - Science and Technology Publications. DOI: 10.5220/0006309705040511
Current student projects
Currently, there are no student projects open.
Completed student projects
Research Project INFOTECH
- Use and Evaluation of Bayesian Networks to Predict the Performance of ML Solutions. Winter Semester 2020/2021
- ACP Dashboard: An interactive visualization tool for selecting Analytics Configurations in an industrial setting. June - December 2017
- Relevance of the two adjusting screws in data analytics: data quality and optimization of algorithms. January - July 2017
Study Project INFOTECH
- Performance Analysis of Machine Learning Models (MLM-Perf). Winter Semester 2019/2020
Practical Course Information Systems
- The Machine Learning Algorithm Wars (ML-Wars). Summer Semester 2019