Embedded Files
Conference Talk
ASME IDETC-CIE 2025
Technique Research Paper :
"Inverse Inference of Aperiodic Structures: Decoding Signed Distance Fields into Learnable Parameters for 3D Reconstruction" by Shaoliang Yang (August 2025, LA, USA)
Abstract: This work presents a learning-based framework for performing inverse inference on complex 3D aperiodic structures represented by their signed distance fields (SDFs) and an automatic dataset generation approach. Since aperiodic structures have unique physical properties and an emerging trend of inverse design, the need to analyze complex structures with machine-learning approaches cannot be ignored. To effectively understand complex 3D geometries represented by their SDFs, we propose an inverse-inference framework to decipher complex SDFs into interpretable key design parameters, notably the periodicity factor and a fill fraction scale parameter, which govern the structure's underlying periodicity properties and spatial intensity. Furthermore, our model predicts target SDFs with complex 3D functions as labels to mathematically encapsulate the input SDF's intricate geometric and topological characteristics. The dataset generation is also based on the key design parameters, enabling direct supervision of model-generated prediction parameters with ground truth key parameters. Our learning-based approach robustly recovers these governing parameters from aperiodic SDF representations. Experimental results show that our model accurately infers these key parameters and precisely reconstructs the SDF with the inferred key parameters. Our model can be used as an analytical tool for inverse engineering of aperiodic structures in fields such as materials science, computer graphics, and design optimization.
CIE Poster Session:
"Automatic Generation of 2D NURBS-based Geometries" by Kevin Wang (August 2025, LA, USA) - Best Poster Award with Travel Stipend!
Abstract: Isogeometric Analysis (IGA) bridges the gap between CAD and FEA by utilizing NURBS to represent complex geometries. However, converting point cloud data into NURBS remains a labor-intensive process. We propose a generative model that automatically maps 2D Cartesian point clouds to NURBS parameters—including control points and weights—to construct high-fidelity NURBS curves. These curves enable efficient and accurate IGA. Our method successfully models various 2D superformula shapes, streamlining the geometry creation pipeline for analysis and facilitating efficient analysis.
12th International Conference on IsoGeometric Analysis (IGA 2024)
Abstract Presentation:
Machine Learning and IGA: "Automatic Generation of NURBS-Based Geometries For IsoGeometric Analysis" by Kevin Wang (October 2024, Augustine, Florida, USA)
Abstract: Isogeometric analysis (IGA) combines computer-aided design (CAD) and finite element analysis (FEA) using Non-Uniform Rational B-Splines (NURBS) for both geometric representation and solution approximation. Despite its benefits like enhanced accuracy, IGA faces challenges in accurately modeling complex geometries and creating corresponding high-quality meshes. This paper introduces an advanced analytical framework designed to streamline and theoretically support the generation of curvilinear configurations. This includes a refined interpolation method for adjusting the coordinates of initial results, enhancements in NURBS surface and knot vector refinement, and a more robust material analytical solution in IGA. This framework effectively closes the gap between generated data and its physical representation. The notable advancements of this approach include the integration of curve generation and a comprehensive framework for a more precise and extensive representation of meshes and material properties.
ASME IDETC-CIE 2024
Technique Research Paper :
"Enhancing Isogeometric Analysis with NURBS-Based Synthesis" by Shaoliang Yang and Kevin Wang (August 2024, DC, USA)
Abstract: Isogeometric analysis (IGA) is a computational technique that integrates computer-aided design (CAD) with finite element analysis (FEA) by employing the same basis functions for both geometry representation and solution approximation. While IGA offers numerous advantages, such as improved accuracy and efficiency, it also presents several challenges related to geometric modeling. Some of these challenges include accurately representing complex geometries with NURBS (Non-Uniform Rational B-Splines) or other basis functions used in IGA and generating high-quality meshes that conform to the complex geometry represented by NURBS curves/surfaces. This paper introduces an analytical framework to provide a more efficient and theoretically grounded method for generating curvilinear configurations and its analytical solution in IGA, bridging the gap between generated data and its physical representations. This innovative approach is distinguished by integrating the NURBS parameterization in curve generation and providing a corresponding framework to achieve a broader and more accurate explanation of meshes and properties, especially constructing new coordinates and calculating the physical displacements under these conditions. Our model enables the analytical understanding of complex curves from the UIUC airfoil and superformula datasets, demonstrating a deeper dive into simulations. This study signifies a pivotal juncture wherein machine-learning-based complex geometrical formulations are synergistically combined with actual isogeometric analysis.
Invited Talk
ASME IDETC-CIE 2025
Guest Speaker:
Early Career Research: Lightning Talks symposium in the 51st Design Automation Conference (DAC) "Practical AI-Enabled Shape Parameterization for Complex Engineering Designs" by Dr. Souma Chowdhury (Summer 2025)
Aerospace Corporation
Guest Speaker:
"Computational and Machine Learning-Driven Next-Generation Engineering Design Tools: Towards Advanced Manufacturing" by Dr. Uchechukwu Agwu (Summer 2025)
SRI International
Guest Speaker:
"Computational and Machine Learning-Driven Next-Generation Engineering Design Tools: Towards Advanced Manufacturing" by Dr. Morad Behandish (Spring 2025)
Santa Clara University
Guest Speaker:
MECH speaker series "Next-Generation Engineering Design Tools for Advanced Manufacturing" by Dr. Hohyun Lee (Winter 2023)
Huazhong University of Science and Technology
Guest Speaker:
"Computational and Machine Learning-Driven Methods for Next-Generation Engineering Design Tools" by Dr. Liang Xia (Fall 2022)
University of California, Merced
Guest Speaker:
SPARK seminar "Engineering Inspired by Nature" by Dr. Sachin Goyal (Spring 2021)
Page updated
Google Sites
Report abuse