01 Research
View all →
Multi-fidelity Modeling
High-fidelity simulations are expensive. We develop surrogate models and super-resolution methods that provide physics-informed insight with reduced computational demands.
Uncertainty Quantification
Sensor limitations and noisy data introduce uncertainty. We use generative deep learning to characterize distributions from sparse data and enhance confidence in predictions.
Foundation Models for Engineering
Labeled engineering data is scarce. We develop self-supervised and pre-trained models that can be fine-tuned for specific tasks without thousands of labeled examples.
