This presentation focuses on the root cause discovery for minor quality in high-variant automotive supply production. With the help of self-learning and explaining AI, complex root-cause findings can be reduced from several days to hours.
High variant complexity is caused by the production of around 11,000 transmissions per day in 700 variants. Every transmission consists of up to 600 parts. There are various suppliers and almost 1,000 machineries in the plant.
Why are semantic networks and thus explanatory AI needed?
The domain experts want to understand why an AI decides how. They do not want to understand the algorithms behind it, but want the reasons to be presented transparently from their domain point of view. On the one hand, this ensures acceptance, but on the other hand, complex disturbing factors are uncovered. From this, domain experts can derive modified procedures or processes.
Why are self-learning algorithms and thus adaptive AI needed?
In such a multi-variant production, the processes are very complex and constantly changing.
Only an AI that permanently brings added value - and not just for a limited period of time - can assert itself here. Therefore, the AI must adapt to the changes. It is only worthwhile for companies to introduce and roll out AI if this AI makes the adjustments itself, i.e. without the intervention of data scientists.
The Speaker of the Tech Talks is Britta Hilt.
Britta Hilt has focused on Artificial Intelligence since 2011, mainly from a usage perspective. She is co-founder and managing director of the AI company IS Predict which has focused on automating Data Science as well as on explaining AI.
Date: June 13, 2023, 10:00 - 11:00 CEST