Integration of multimodal data and explainable artificial intelligence for root cause analysis in manufacturing processes
Matteo Calaon, Tingting Chen, Guido Tosello
Abstract
Nowadays, the growing complexities of manufacturing processes and systems make it difficult to identify the root causes of critical deviations in performance. Conventional methods often fall short in capturing the multifaceted nature of these challenges, despite a wealth of diverse untapped manufacturing data. To harness the full potential of diverse data sets and transform them into a valuable asset to guide root cause exploration, this paper presents an innovative approach that combines multimodal predictive analysis and explainable artificial intelligence (XAI) to uncover insights into system dynamics. This work contributes to a paradigm shift in industrial decision-making regarding manufacturing diagnostics.