课程题目：Robust Geometry Processing for Physical Simulation
授课讲者：Zhongshi Jiang (姜仲石), Meta
课程摘要：Partial differential equations (PDEs) play a crucial role in various fields such as scientific computing, computer graphics, and engineering for simulating physical phenomena. It is desirable for a PDE solver to be user-friendly, where the user specifies the domain boundary, boundary conditions, and governing equations as input, and the solver provides an evaluator that can calculate the solution's value at any point within the domain. However, despite significant research efforts and widespread interest from academia and industry, existing open-source or commercial software often falls short of this ideal. One of the main reasons behind this deficiency lies in the limitations and lack of robustness in the geometry processing algorithms used to convert raw geometric data into a format suitable for PDE solvers. In this discussion, I will highlight the current state-of-the-art's limitations and propose an integrated pipeline that addresses these challenges comprehensively. This integrated approach considers the entire process, including data acquisition, meshing, basis design, and numerical optimization, as a unified challenge. By optimizing and automating different phases of the pipeline, we can achieve increased efficiency and automation while balancing tradeoffs between various stages. Furthermore, it introduces intriguing new geometry processing challenges.
讲者简介：Zhongshi is an AI Research Scientist at Reality Labs, Meta. He finished PhD degree in New York University. Zhongshi’s research aims to build robust algorithm and tools to study geometry representations, with applications in mesh generation, robust simulation, scientific computing, and digital fabrication. Some of his research has been recognized by Adobe Research Fellowship 2018, Geometry Processing Dataset Award 2019, and Jacob T. Schwartz PhD Fellowship 2020.
课程摘要：Contact is ubiquitous and often unavoidable, and yet modeling contacting systems continues to stretch the limits of available computational tools. In part, this is due to the unique hurdles posed by contact problems. Several intricately intertwined physical and geometric factors make contact computations hard, especially in the presence of friction and nonlinear elasticity. I will discuss our work on optimization-based finite element solver, which is constructed for mesh-based discretizations of nonlinear elastodynamic problems supporting large nonlinear deformations, implicit time-stepping with contact and friction. Built on top of a smooth barrier reformulation and a custom Newton-type optimization, it is a first-of-its-kind "plug-and-play" contact simulation framework that provides convergent and unconditionally feasible intersection-free trajectories. The scheme also enables future studies of differentiable simulations of nonsmooth physics-constrained inverse problems in design, control, and robotics.
讲者简介：Chenfanfu Jiang is an associate professor of Mathematics at UCLA. He was a recipient of the UCLA Edward K. Rice Outstanding Doctoral Student Award (2015), NSF CRII award (2018) and NSF CAREER award (2020). He directs UCLA Multi-Physics Lagrangian-Eulerian Simulations (MultiPLES) Laboratory with research projects spanning scientific computing, computer graphics, computational mechanics, machine learning, and robotics.
讲者简介：Evan Y. Peng is an Assistant Professor in the University of Hong Kong Electrical and Electronic Engineering and Computer Science . Before joining HKU, he was a Postdoctoral Research Scholar in the Stanford University Computational Imaging Laboratory. He received my PhD in Computer Science from the Imager Lab, the University of British Columbia. During the PhD, He was a Visiting Student Researcher at Visual Computing Center, KAUST, and at Stanford University. He received both my MSc and BS in Optical Science and Engineering from State Key Lab of Modern Optical Instrumentation, Zhejiang University.
His research interest lies in the interdisciplinary field of Optics, Graphics, Vision, and Artificial Intelligence, particularly with the focus of: Computational Optics, Sensing, and Display; Holographic Imaging/Display & VR/AR/MR; Computational Microscope Imaging; Low-level Computer Vision; Inverse Rendering; Human-centered Visual & Sensory Systems.
课程题目：Computational Assemblies: Bridging the Gap between Design and Production
授课讲者：Ziqi Wang (汪子琦), ETH Zurich
课程摘要：An assembly refers to a collection of parts joined together to achieve a specific form and/or functionality. Assembling enables the creation of large, intricate structures or machines from smaller, simpler parts. The ability to repeatedly assemble and disassemble parts greatly enhances efficiency in transportation and maintenance. Due to these advantages, assemblies are ubiquitous in our daily lives, including most furniture, household appliances, and buildings.
The recent advancements in manufacturing have introduced numerous digital fabrication techniques (e.g., 3D printing), which have significantly simplified the production of complex, monolithic objects. However, designing assemblies remains an onerous task due to the lengthy process from design to production. Many projects often disregard considerations for structural integrity, manufacturability, and construction constraints at their early design stages. This leads to labor-intensive redesigns and iterations, elongating the process and increasing the cost.
This course aims to mitigate these challenges by integrating a variety of computational design algorithms into the design stage. These include computational geometry modeling, structural analysis, and assembly planning. By adopting these approaches, designers can concentrate more on high-level design decisions, consequently driving digital transformation in the architecture, construction, and manufacturing industries.
讲者简介：Ziqi Wang is a Postdoctoral researcher at Computational Robotics Lab (CRL), ETH Zurich, under the supervision of Prof. Dr. Stelian Coros. He completed his Ph.D. at the Geometric Computing Laboratory (GCM) at EPFL, guided by Prof. Dr. Mark Pauly. Prior to his Ph.D., he earned his Bachelor's degree in Mathematics in 2017 from the University of Science and Technology of China (USTC).
Since 2018, his research has been supported by NCCR Digital Fabrication, a research initiative aimed at revolutionizing architecture and construction through the integration of digital technologies. His work has been published in top-tier graphics journals and conferences and has been utilized in actual architectural projects. His current research is centered on computational assemblies, with the objective of designing, analyzing, and fabricating complex assemblies using computational techniques and robotics.
课程题目：Planar spline curve design from research to engineering
授课讲者：Zhipei Yan, Nvidia
课程摘要：In this course we will talk about some fundamental math knowledge about Computer-Aided Geometric Design like polynomial curves/surfaces, curvature, continuity, etc. Then we will talk about some typical shape modeling tools and digital content creation tools. And then we can talk about some engineering related topics when utilizing research techniques in real practical problems.
讲者简介：Dr. Zhipei Yan is a Sr. Software Engineer at Nvidia Inc. He received his Ph.D. in Computer Science from Texas A&M University. Before that, he received his B.Sc. in Mathematics from University of Science and Technology of China. His main research interests are computer graphics, geometric modeling, curves and surfaces, etc.
课程题目：Acquiring Stylized Motor Skills for Physics-based Characters
授课讲者：杨泽世, 米哈游
课程摘要：Character animation is a core research topic in computer graphics, which studies synthesizing realistic movements for virtual characters. Physics-based character animation utilizes physics simulation to generate physically plausible character motions. One long-standing goal of physics-based character animation is to equip simulated characters with vast and agile motor skills. Most recent physics-based animation methods learn these agile and impressive motion skills by imitating motion capture data or human demonstrations. However, most of these methods mainly focus on motion tracking, and cannot discover novel skills that are visually fundamentally different from reference motions. Therefore, these imitation-based methods cannot be applied to motor skill learning tasks where high-quality motion capture data is not available. In this seminar, I will provide a concise overview of the key concepts in physics-based character animation, including motion control and deep reinforcement learning. Then I will present several lines of work that develop diverse motor abilities, such as high jumping and hand manipulations, for physics-based characters without using task-specific motion data. Finally, I will explore intriguing studies that demonstrate the broader applicability of computational models commonly used in character animation, such as deep reinforcement learning, in solving other problems within the realm of computer graphics.
讲者简介：杨泽世，米哈游图形学研究员。2018年在中国科学技术大学获得物理学学士学位， 在2023年在Simon Fraser University获得计算机博士学位，在2022-2023年期间在北京大学可视计算实验室访问交流。研究兴趣包括但不限于：基于物理的角色动画，机器人，以及强化学习。在SIGGRAPH， Eurographics, Interactive 3D Graphics上发表文章多篇。
课程摘要：Generative AI has made remarkable progress in recent few years, among which text-guided content generation is one of the most practical techniques since it enables natural interactions between human instruction and AIGC. Text-to-3D has become an emerging yet highly active research field due to the development in text-to-image as well 3D modeling technologies such as NeRF. This course will focus on the task of text-driven 3D scene generation. It would provide a comprehensive summarization about how text-to-3D technology is used in scene generation and would help students quickly catch up with the progress in the research community.