NSF International Research Experiences for Students (IRES) Project
NSF IRES: Physical AI Design

JAIST-NSF IRES Physical AI Workshop

2023/8/18(Fri) 9:30-12:30 Tokyo Time, JAIST+Online Hybrid Workshop

主催者Organizers: Prof. Haoran Xie/謝浩然, 北陸先端科学技術大学院大学(JAIST); Prof. Bo Zhu, Georgia Institute of Technology; Prof. Takeo Igarashi/五十嵐健夫, 東京大学(University of Tokyo)

研究予算Funds:NSF International Research Experiences for Students (IRES), JAIST Research Grants for JAIST Science Hub 令和5年度研究拠点形成支援事業.

開催場所Place: 北陸先端科学技術大学院大学 知識棟4階K42 コラボレーションルーム1
オンライン参加Online Particiaption (zoom url): URL


Introduction


Program: 8/18(Fri) 9:30-12:30 in JST Time


    ----開会開催Opening----
  • 9:30 AM - 9:40 AM Opening by Prof. Haoran Xie (JAIST)

  • ----招待講演Invited Talk----
  • 9:40 AM - 10:30 AM invited talk (online), Prof. Bo Zhu (Georgia Institute of Technology)
  •           Computing in Miniature: From Graphics to Science
    ----学生発表Student Presentation----
  • 10:40 AM - 11:00 AM Ryan Zesch (Texas A&M University)
  •           Neural Collisions and NeRF Garments
  • 11:00 AM - 11:20 AM Humphrey Yang (Carnegie Mellon University)
  •           Generative Design and Synthesis of Transmittive Compliant Mechanisms
  • 11:20 AM - 11:40 AM Alina Chadwick (Dartmouth College)
  •           MicroGlam: Microscopic Skin Image Dataset with Cosmetics
    ----休憩時間Rest----
  • 11:50 AM - 12:10 AM Mikiya Kusunoki (JAIST)
  •           Rapid Fabrication of Soft Mechanisms Inspired By Origami Twisted Tower
  • 12:10 AM - 12:30 AM Tianyu Zhang (JAIST)
  •           Sketch-Guided Two-Stage Text-to-Image Generation with Spatial Control
    ----閉会開催Closing----

招待講演 / Invited Talk

    Title:

    Computing in Miniature: From Graphics to Science

    Abstract:

    Step into a magnificent miniature wonderland where small insects gracefully traverse the surface of the pond, raindrops delicately splash upon vibrating foliage, and dragonflies glide and soar through the air. Here, a symphony of small-scale interactions unfolds, captivating the imagination. In this talk, I will showcase our group's recent progress in developing advanced computational tools tailored to investigate the intricacies of these fascinating systems. Our work encompasses three key aspects. First, we invent novel geometric representations and solvers to simulate the complex processes of small-scale solids and fluids. Second, we construct computational design and optimization algorithms to push the performance limits of these systems and synthesize designs inspired by nature's ingenuity. Lastly, we develop machine learning algorithms that integrate physical principles to uncover the governing mechanics underlying these intricate systems. These advancements constitute a fully automated computational pipeline, enabling the exploration of complex systems at a small length scale. I will showcase applications of these computational tools in different disciplines, including pathogen transmission understanding, miniature robot design, new material discovery, and turbulence dynamics prediction.

    Bio:

    Bo Zhu is an Assistant Professor affiliated to the School of Interactive Computing at Georgia Tech. His research focuses on the intersection of computer graphics, computational physics, computational design and fabrication, and scientific machine learning. He obtained his Ph.D. in Computer Science from Stanford University and completed postdoctoral training at MIT CSAIL. His works have been supported by multiple funding agencies and industrial sponsors. He is a recipient of the NSF CAREER Award in 2022.