Speaker

Hybrid AI Architectures for Intelligent Diagnostic and Predictive Applications

A powerful generic AI workbench – the Raptor Suite – is presented which allows the complete development of automatable intelligent applications for reactive and predictive diagnosis with the ability to learn automatically. The Raptor technology contains all ingredients which are needed to build such applications on a very high and sophisticated level. The applications are modelled as knowledge bases – and not programmed!
These applications rely on a hybrid AI architecture. Hybrid AI systems are programs that not only have learnt from big data to solve a problem. Hybrid systems also take advantage from explicit domain and expert knowledge. They combine the capabilities of symbolic knowledge processing with the power of subsymbolic machine learning and data science. This approach makes AI applications more powerful, explainable, comprehensible and last but not least! auditable.
The Raptor technology is applicable to any kind of engineered systems and processes. It will be demonstrated at the example of complex and challenging applications from semiconductor industry as 'Defect Density Engineering', 'Deviation Analysis & Management' and 'Ion Implantation.