Our Research

In the Design Innovation & Generative InTelligence (DIGIT) Lab, we develop advanced AI and machine learning (ML) methods for engineering design, targeting some of today’s most challenging problems in design ideation, design automation, and design for manufacturing.

• Design ideation: Human creativity is constrained by individual knowledge and experience, with outcomes often highly uncertain. We explore how AI can enhance the generation of creative design ideas, fostering human-AI co-creativity and accelerating design innovation.
• Design automation: We develop AI and ML methods to enhance the generation of detailed design solutions, addressing the challenges of vast design spaces, costly design evaluations (e.g., simulations and experiments), and inherent uncertainties.
• Design for manufacturing: We leverage AI and ML to design better manufacturing processes, enable new manufacturing techniques, and optimize design solutions by accounting for manufacturing imperfections.

Engineering design problems span a wide range of domain-specific contexts–such as aerospace (e.g., designing aerodynamic shapes and structures), materials science (e.g., designing new materials and metamaterials), and robotics (e.g., designing soft actuators). Our research abstracts these domain-specific problems and reformulates them as mathematical problems, developing AI-driven design methodologies to solve them while incorporating domain knowledge when needed.

Our methodologies usually follow a “generative” paradigm, where AI learns to generate design solutions by modeling their underlying distributions and capturing uncertainties. Our goal is to leverage AI and ML to improve efficiency, stimulate creativity, and uncover new knowledge and insights in design processes, while ensuring the generalizability and trustworthiness of these methods.

For more information, you can explore our paper collection or check out our open-source code.