Simulation-based machine learning – Trains machine learning models in simulated environments while bridging the “reality gap” to ensure robust real-world performance.
Predictive maintenance – Develops models for anomaly and out-of-distribution detection to monitor machine health and reduce unscheduled downtime.
Adaptive deep learning – Focuses on efficient online, continual, and distributed learning, allowing models to make real-time adjustments on live production streams.
Neuromorphic computing – Employs brain-inspired neuromorphic sensors and chips to minimize energy consumption while maximizing computational throughput.
Trustworthy deep learning – Develops novel methods to enhance the explainability, accountability, and overall trustworthiness of deep learning systems in high-tech manufacturing.