As an FZI scientist, he deals with the long-term robustness of artificial intelligence (AI). Under the slogan “I think my AI is broken”, he tackles the problems that sooner or later happen exactly where it hurts: in the productive environment of industrial companies. Due to frequent changes in real-life circumstances and the data that represents and describes them, the catastrophic failure of AI is often inevitable without appropriate strategies in place.
As the FZI's Intelligent Systems and Production Engineering division, we have been dealing with this problem for several years as the subject of numerous industrial research projects. We always work in close cooperation with our industrial partners to implement real solutions and achieve genuine success.
Is your AI sufficiently robust? Let's talk about it!
Keywords: Machine Learning Operations (MLOps), Incremental Learning, Life-Long Learning, Data Stream Mining