Overview
"Scientists study the world as it is; engineers create the world that never has been."
- Theodore von Kármán
Welcome to the Computational Intelligence and Multiphysics Simulation (CIMS) Lab at the University of Minnesota! We are a multidisciplinary research team that leverages Computational Mechanics, Scientific Computing, and Artificial Intelligence (AI) to address resilience and sustainability challenges related to materials, structures, and geosystems under extreme conditions. Our research advances data-enabled computational methods, including meshfree methods, reduced-order modeling, physics-informed machine learning, and data assimilation, for the simulation, optimization, and control of complex multiscale and multiphysics systems.
Our research is centered on three core themes:
- Differentiable and AI-Enhanced Scientific Computing
- Multiscale Mechanics of Geological, Composite, and Energy Materials
- Digital Twins and Inverse Modeling for Multiphysics Processes in Subsurface Science and Extreme Hazard Events
For more information about our active research, see Research and Publication.
Recent News
- 08/2026 Awards: Congratulations to the CIMS Lab on leading a project selected for the DOE Genesis Mission award, in collaboration with PNNL and LBNL! We are also honored to participate in another Genesis Mission project led by Prof. Peter Kang. [UMN News] [CSE News]
- 08/2026 Welcome: We are delighted to welcome Hsiang-Liang (Joseph) Lin, a new CEGE PhD student, to the CIMS Lab! Hsiang-Liang joins us with a CEGE Department Fellowship. Congratulations!
- 06/2026 Awards: Honored to receive the NSF CAREER Award!
- 06/2026 Awards: Congratulations to Honghui on winning Third Place in the Student Research Competition at the 60th U.S. Rock Mechanics / Geomechanics Symposium (ARMA 2026) for his research on differentiable computational geomechanics.
- 05/2026 Short Course: Profs. WaiChing Sun, Nick Vlassis, JS Chen, and I will teach the fifth version of our short course Machine Learning for Solid Mechanics at WCCM-ECCOMAS 2026 Munich on July 19th, July 2016. We will cover selected topics in generative AI, LLMs, geometric learning, and the use of coding agents (Claude, Codex, and other open-source alternatives) for computational mechanics. Due to space constraints, we can only accept 40 registrations. The registration deadline is June 10, 2026. Please feel free to share this announcement with students, colleagues, and others who may be interested. [Link]
- 05/2026 Awards: The CIMS group has been awarded a Environment and Natural Resources Trust Fund (ENRTF) grant to develop machine learning-enabled tools for identifying complex subsurface fracture networks and improving groundwater flow prediction. We gratefully acknowledge the support of the ENRTF and the LCCMR.
- 05/2026 News: Congratulations to Binyao on successfully passing the Preliminary Oral Exam—excited to see you move on to the next milestone.
- 04/2026 Events: Dr. He delivered an ISRM AI Café Talk titled “Hybrid ML–Physics Forward and Inverse Modeling of Geomechanics and Geophysical Flow Hazards,” organized by the International Society for Rock Mechanics and Rock Engineering. Thanks for the hosting by Hongkyu and Lina, and the invitation from Wei; It was a productive discussion.
- 04/2026 Publications: Congratulations to Honghui on the acceptance of the paper “JAX-MPM: A Learning-Augmented Differentiable Meshfree Framework for GPU-Accelerated Lagrangian Simulation and Geophysical Inverse Modeling” for publication in Engineering with Computers [Link].
- 03/2026 Awards: Our CIMS group receives a National Security Research Institute (NSRI) Seed Grant to advance generative Bayesian learning for complex flow dynamics, in collaboration with the Stochastic Hypersonics Research Group led by Prof. del Val (AEM). We sincerely thank NSRI for their support.
- 02/2026 Events: We are pleased to invite submissions to our mini-symposium:
- "MS 208 - Data-Driven Approaches for Solid Mechanics" at the 20th U.S. National Congress of Theoretical and Applied Mechanics (USNC-TAM 2026), June 21–25, 2026, Pasadena, California
- "MS426 – Data-driven Approaches in Mechanics" at the 17th World Congress on Computational Mechanics (WCCM), 19 - 24 July 2026, Munich, Germany.
- 01/2026 News: Congratulations to Zihan on passing the Preliminary Written Exam!
- 01/2026 Events: Dr. He will serve as a co-chair of Session IS01 "AI/ML Applications in Rock Mechanics" at the 60th U.S. Rock Mechanics / Geomechanics Symposium of the American Rock Mechanics Association (ARMA), to be held in Tucson, Arizona, USA, on June 21–24, 2026.
- 01/2026 Research: The CIMS Lab has completed an exciting MnDOT-funded study demonstrating that onboard electric-vehicle data can assess pavement conditions with up to 94% accuracy. The project was a wonderful collaboration with Prof. Mihai Marasteanu (PI), Prof. Raphael Stern, Research Scientist Mugurel Turos, and an outstanding team of students. Related news: [Link] [Research Summary] [Link].
- 12/2025 Outreach: Dr. He gave a webinar talk for the Center for Transportation Studies (CTS) on "Preparing for AI use in transportation".
- 12/2025 Publications: New paper on arXiv: Differentiable Inverse Modeling with Physics-Constrained Latent Diffusion for Heterogeneous Subsurface Parameter Fields (arXiv:2512.22421, Dec. 27, 2025)
- 12/2025 News: Congratulations to Honghui on passing the Preliminary Oral Exam!
- 10/2025 News: Dr. He has been appointed as a member of the Editorial Board of Acta Mechanica Sinica.
- 09/2025 Events: Dr. He presented “A Differentiable Meshfree Framework for Geomechanics and Natural Hazard Modeling” at the inaugural Artificial Intelligence and Digital Twins for Earth Systems (AIDT4ES) Workshop, co-hosted by the Oden Institute for Computational Engineering and Sciences and the U.S. Association for Computational Mechanics (USACM) Energy & Earth Systems (EE&S) Technical Thrust Area.
- 09/2025 Welcome: Welcome Niketha (MS student in Computer Science) to the CIMS Lab!
- 09/2025 Publications: Our collaborative paper with PNNL, titled "Integrating Physics-Informed and Data-Driven Neural Networks into Earth System Models: A Comparative Study for Compound Flood Simulation at River-Ocean Interfaces" has been accepted for publication accepted by Journal of Geophysical Research: Machine Learning and Computation. Congratulations to the team!
- 09/2025 Publications: Binyao's paper "History-Aware Neural Operator (HANO)" has been accepted for publication in Computer Methods in Applied Mechanics and Engineering. This study introduces and verifies a novel idea for robust data-driven modeling of generic path-dependent materials (e.g., geomaterials and damaged alloys) through learning from historical loading data while bypassing the need for constructing unphysical internal variable.