Research Topic

Physics-Guided AI for Engineering Simulation & Design

We integrate physical understanding with reliable AI workflowsto speed up simulations,
predict future physical behavior, and guide better design decisions.

Research Overview
Physics-guided AI roadmap graphic
Key Components of AI-based Design
1

Simulation

Computer simulations provide physics-consistent data that anchor AI models to real engineering behavior.

2

AI Modeling

AI models learn how physical systems behave by capturing patterns and rules, not just fitting data points.

3

AI Prediction

Trained AI models forecast how systems evolve over time, turning simulation knowledge into usable future insight.

4

Design Optimization

AI connects prediction to decision-making by searching for better designs under physical and practical constraints.

Our Signature AI Techniques for AI-based Design
1

Data-Independent AI

AI remains reliable even with limited, noisy, or inconsistent data, so its performance does not depend on a specific setup.

2

Long-Term Forecasting

AI predicts future physical behavior over long time horizons while staying consistent with physical trends.

3

Uncertainty Quantification

AI quantifies uncertainty to inform how much confidence engineers should have in each prediction.

4

Generative Design

AI suggests numerous novel design candidates, enabling efficient discovery of high-performing designs.