The manufacturing sector is a worldwide leader in application of AI. Designing better products, cutting unplanned downtime, keeping employees safe - all of these while adapting to the ever-changing market can be challenging. Artificial Intelligence is a core element of the Industry 4.0 revolution and can give management an opportunity to move from a reactionary to a strategic mindset.
The conveyor belt is a well-oiled machine, but small mistakes and errors can easily slip in. If these errors are not spotted and corrected early, the downstream costs of fixing the original problem or settling claims tend to be orders of magnitude bigger.
Our challenge was to develop a model that improves performance in anomaly detection and enhances the human knowledge with precise early detection.
Our solution identifies which parts of the assembled product are faulty and need further inspection at a very early stage, preventing headaches down the line. We built this tool for a large manufacturer of engines for earthmoving machines, that was suffering from a high claims rate for faulty machines.