Header

School of Engineering

Rieter Endowed Professorship for Industrial Artificial Intelligence

"We focus on developing learning-based control and optimization methods that integrate expert knowledge and symbolic relationships to optimize industrial manufacturing processes, enhance autonomy, and ensure safety in the face of complex dynamics and variability."

Dr. Alisa Rupenyan, Rieter Endowed Professorship for Industrial Artificial Intelligence

The Rieter endowed professorship at the Centre for Artificial Intelligence has a particular focus on the development and application of principled, neuro-symbolically inspired learning in conjunction with processes in production and service. This combines machine learning with knowledge-based systems. 

Fields of Expertise

  • Learning-based control and optimization of manufacturing systems 
  • Neuro-symbolic AI for modeling and optimization of manufacturing processes
  • Expert knowledge integration in machine learning models

Industrial manufacturing processes pose various challenges to their optimization and control, due to their complex dynamics, lack of observability, and large variability. Apart of finding the right optimization approaches and parameters, finding the right strategy to adapt to different requirements, different components of the equipment, or changes in the environment is crucial. Leveraging existing information in the form of graph-based contextual information, logical rules, or another abstracted form of pre-existing knowledge is very promising to alleviate the large variability challenge. The research focus of the Industrial AI (IAI) group is on the development of learning-based control and optimization methodologies, incorporating the existing expert knowledge, and symbolic relationships in the process, tapping into the newly emerging neuro-symbolic AI approaches. Our goal is to demonstrate that industrial processes can be optimized, combining machine learning models (such as neural networks) and symbolic reasoning (graph-based representations, logical rules), and methods from control and sequential decision making, as well as domain expertise. We are interested in achieving increased autonomy for systems in various domains (manufacturing, construction, process industries), and working with distributed systems, while striking a balance between autonomy and human oversight. By developing innovative solutions that leverage both quantitative data and qualitative domain understanding, we contribute to more adaptive, efficient, and safe industrial processes across various domains.

Services

Team

Projekte

Publikationen