Machine Learning
Finds patterns, trends and fast predictions in engineering and operational data.
What this expert provides
Finds patterns, trends and fast predictions in engineering and operational data, and always checks ML results against physics before they reach a design or operating decision. Typical outputs include surrogate models for CFD/FEA, physics-informed neural networks (PINNs), and design optimization sets.
Expert Profile
AI-PNA uses machine learning to find patterns and trends. Capabilities include Surrogate models (Gaussian process, gradient boosting, neural network), Physics-informed neural networks (PINNs), Neural operators for field predictions, and Design optimisation (NSGA-II/III). It normalises performance against a physics baseline, detects change points, and forecasts ahead with confidence intervals.
Key Capabilities
- Surrogate modeling (Gaussian process, gradient boosting)
- Physics-informed neural networks (PINNs)
- Neural operators and graph surrogates for field predictions
- Design optimisation (NSGA-II/III and Bayesian)
- Anomaly detection and condition monitoring
- Reduced-order and data-driven dynamics (SINDy, DMD)