DeCoDe Lab

DeCoDE Lab is based in the Mechanical Engineering department at MIT. Our vision is to create a world where humans and AI design together to solve our biggest challenges. We study machine learning and optimization methods to better design complex mechanical systems and assist teams of human designers in creating better products.

Our work lies at the intersections of Mechanical Engineering, Artificial Intelligence, and Human-Computer Interaction. A few questions we are interested in: How can algorithms synthesize high-performing designs that meet real-world requirements? How can algorithms help discover or create creative designs that have never been seen before? How can we enable distributed teams of people to create better products? How can we design, develop, and deploy advanced engineering material systems for complex non-linear inverse problems? How can we quickly and reliably evaluate thousands of ideas to accelerate innovation? While these questions address different areas, our underlying approach is to mathematically characterize these questions as generalizable machine learning and optimization problems, make testable predictions for new problems, and tie together or understand individual empirical results that researchers have generated.

Our techniques apply to a wide range of problems in Engineering. We aim to transform the way humans design products and measure our success by the impact of our work on society. We believe in reproducible and open-source science and do our part by making most of our research code and papers available online.

Selected Publications [See all]

Song, B., Miller, S., Ahmed, F.
Attention-enhanced Multimodal Learning For Conceptual Design Evaluations
Under review
PDF

Heyrani Nobari, A., Srivastava, A., Gutfreund, D., Ahmed, F.
LINKS: A dataset of a hundred million planar linkage mechanisms for data-driven kinematic design
In IDETC 2022
PDF Project Page Web Link Dataset

Regenwetter, L., Ahmed, F.
Design Target Achievement Index: A Differentiable Metric to Enhance Deep Generative Models in Multi-Objective Inverse Design
In IDETC 2022
PDF Project Page Web Link

Song, B., Miller, S., Ahmed, F.
Hey, AI! Can You See What I See? Multimodal Transfer Learning-based Design Metric Prediction for Sketches with Textual Descriptions
In IDETC 2022
PDF Web Link

Bagazinski, N. J., Ahmed, F.
Ship Deck Object Placement Optimization Using a Many-Objective Bilevel Approach
In IDETC 2022
PDF Web Link

Regenwetter, L., Heyrani Nobari, A., Ahmed, F.
Deep Generative Models in Engineering Design: A Review
In Journal of Mechanical Design 2022
PDF Web Link Dataset

Chen, W., Ahmed, F.
MO-PaDGAN: Reparameterizing Engineering Designs for Augmented Multi-objective Optimization
In Applied Soft Computing 2021
PDF Project Page Web Link

Heyrani Nobari, A., Chen, W., Ahmed, F.
PcDGAN: A Continuous Conditional Diverse Generative Adversarial Network For Inverse Design
In ACM KDD 2021
PDF Project Page Web Link

Announcements

We have an available PostDoc position in deep generative models and geometric deep learning. Apply here.

Prospective Ph.D. students - If you are interested in joining the DeCoDE lab, you can apply to the Computational Science and Engineering Program or to the Mechanical Engineering department at MIT.

If you are interested in joining the DeCoDE lab, drop me an email with the following text in your subject line "Join DeCoDE:" followed by the position you are interested in (for example, a postdoc, intern, visiting student, etc.).

Datasets

A list of engineering design datasets curated by the DeCoDE lab.

LINKS dataset - A dataset of 100 million planar linkage mechanisms and 1.1 billion coupler curves obtained from kinematic simulations. The dataset contains curated curves, another 100 million negative samples, and publicly available simulation software.

BIKED dataset - A dataset of 4500 community-designed bicycles in tabular and image format, along with images corresponding to different bike parts for each bicycle.

FRAMED dataset - A dataset of 4500 bicycle frames and ten performance metrics obtained from physical simulations.

Aircraft dataset - A dataset of 4045 3D aircraft models from Shapenet and their lift/drag performance values.

Airfoil dataset - A synthetic dataset of 48,503 airfoils and their aerodynamic performance computed using OpenFOAM.

Principal Investigator

Faez Ahmed

d'Arbeloff Career Development Assistant Professor
Department of Mechanical Engineering
Massachusetts Institute of Technology
Email: faez at mit dot edu

Faez Ahmed is the d'Arbeloff career development assistant professor in the Department of Mechanical Engineering at MIT, where he leads the Design Computation and Digital Engineering (DeCoDE) lab. His research focuses on developing new machine learning and optimization methods to study complex engineering design problems. Before joining MIT, Ahmed was a postdoctoral fellow at Northwestern University and completed his Ph.D. in mechanical engineering at the University of Maryland. He also worked in the railway and mining industry in Australia, where he pioneered data-driven predictive maintenance and renewal planning efforts.