AIR Seminar: Minghao Guo, PhD student, MIT CSAIL

Speaker: Minghao Guo is a 5th-year Ph.D. student at MIT CSAIL, advised by Prof. Wojciech Matusik.

Talk Title:Neural Generative Design: Generative Models Grounding with Hollywood Physics

Abstract:Generative foundation models have reshaped modern design workflows in science and engineering, yet they often propose designs that violate the very constraints required for real‑world realization. The root cause is objective misalignment: beyond novelty and variety, scientific and engineering objectives must also meet user requirements under the hard constraints of physical reality. In this talk, I will present a generative design workflow that treats physical constraints as first‑class citizens. Drawing on classical computer graphics and scientific computing, I will show how physics‑aware geometric primitives and differentiable constraint solvers couple tightly with generative models to guarantee feasibility by construction, yielding empirically faster and more physically faithful generation across scales -- from macro-scale engineered shapes to nano-scale protein structures.

Bio::Minghao Guo is a 5th-year Ph.D. student at MIT CSAIL, advised by Prof. Wojciech Matusik. He received his master's degree from the Chinese University of Hong Kong and his bachelor's degree from Tsinghua University. Minghao’s research lies at the intersection of computer graphics, generative modeling, and scientific discovery. His research builds generative design systems that translate scientific and engineering intent into designs that are physically valid, manufacturable, and reliable. He is a recipient of the MathWorks Fellowship and the Meshy AI Fellowship, and a finalist for the Roblox Fellowship.