Session: DAC-02-01-Artificial Intelligence and Machine Learning for Challenging Real-World Problems in Design Automation

Paper Number: 88049

88049 - Material Prediction for Design Automation Using Graph Representation Learning 

Successful material selection is critical in designing and manufacturing products for design automation. Designers leverage their knowledge and experience to create high-quality designs by selecting the most appropriate materials through performance, manufacturability, and sustainability evaluation. Intelligent tools can help designers with varying expertise knowledge by providing recommendations learned from prior designs. To enable this, we introduce a graph representation learning framework that supports material prediction of bodies in assemblies. We formulate the material selection task as a node-level prediction task over the assembly graph representation of CAD models and tackle it using Graph Neural Networks (GNNs). Evaluations over three experimental protocols performed on the Fusion 360 Gallery dataset indicate the feasibility of our approach, achieving a 0.75 top-3 micro-F₁ score. The proposed framework can scale up to large datasets and incorporate designers' knowledge into the learning process. These capabilities allow the framework to serve as a recommendation system for design automation and a baseline for future work, narrowing the gap between human designers and intelligent design agents. While our results demonstrate the feasibility of leveraging graph representation learning for feature predictions ongraphically represented CAD models, the configuration of the three experiments shows promise in supporting human-in-the-loop design automation applications.

Presenting Author: Bingbing Li California State University Northridge

Presenting Author Biography: Dr. Li serves as the Associate Director of NASA Autonomy Research Center for STEAHM (ARCS), Assistant Director of DOE Industrial Assessment Center (IAC) at UC Irvine and CSUN (Satellite), Director of Laboratory for Sustainable and Additive Manufacturing. Dr. Li is also the Affiliate Faculty at the Terasaki Institute for Biomedical Innovation.<br/>Dr. Li conducts research in Additive Manufacturing (Metal AM, 3D Bioprinting, Design for AM), AI-powered Design and Manufacturing (Smart Connected Worker, Digital Twins, AR/VR for Manufacturing, AI-assisted Knowledge Graph Design), and Sustainable Manufacturing (Sustainability Analysis, Energy Efficiency, Life Cycle Assessment, Remanufacturing). He teaches undergraduate and graduate courses in the Manufacturing Systems Engineering program.<br/>Dr. Li is one of the Faculty Mentors of the NASA funded Autonomy Research Center for STEAHM (ARCS), NIH funded Building Infrastructure Leading to Diversity (BUILD) Promoting Opportunities for Diversity in Education and Research (PODER), and USDE funded HSI-STEM/ AIMS2 (Attract, Inspire, Mentor and Support Students). He is the Faculty Advisor of the Society of Manufacturing Engineers (SME) Student Chapter S327, and CSUN Chinese Students and Scholars Association (CSSA).<br/>Dr. Li is the member of the Society of Manufacturing Engineers (SME), American Society of Mechanical Engineers (ASME), and American Society for Engineering Education (ASEE).

Authors:

Shijie Bian University of California Los Angeles
Daniele Grandi Autodesk Inc
Kaveh Hassani Autodesk Inc
Elliot Sadler California State University Northridge
Bodia Borijin University of California San Diego
Axel Fernandes California State University Northridge
Andrew Wang Portola High School
Thomas Lu California Institute of Technology
Richard Otis California Institute of Technology
Nhut Ho California State University Northridge
Bingbing Li California State University Northridge