The presentation will discuss the use of AI in anomaly detection in large-scale testing data: powerful neural networks flagging deviations through high reconstruction errors; temporal convolutional networks processing multivariate sensor dependencies across high-speed telemetry streams; and physics-informed neural networks embedding strict thermodynamic governing laws directly into the model loss function, ensuring data-driven models respect physical reality during aggressive transient engine testing operations successfully and reliably across every engineering project.

• Types of anomaly detection techniques which can be used on test bench data
• Acclerating/automating the testing of automotive systems
• Integration of AI and ML in automotive testing