About

Elizabeth Holm is the Richard F. and Eleanor A. Towner Professor of Engineering and Department Chair at the University of Michigan's Department of Materials Science and Engineering. She holds a Ph.D. in Materials Science and Engineering and Scientific Computing from the University of Michigan, as well as an S.M. in Ceramics from MIT. Her research employs computational materials science tools to investigate a variety of materials systems and phenomena. Her areas of focus include the theory and modeling of microstructural evolution in complex polycrystals, the physical and mechanical response of microstructures, the mechanical properties of carbon nanotube networks, atomic-scale properties of internal interfaces, and the application of machine vision and machine learning to microstructural classification and prediction of rare events. Her computational techniques span from atomic-scale molecular dynamics to mesoscale methods such as Monte Carlo, phase field, and cellular automata, as well as continuum-scale finite element analysis. A key aspect of her work involves identifying and developing concepts from data science, including machine learning, machine vision, evolutionary computing, and network analysis, to address materials science questions. Prior to her current role, she served as a professor at Carnegie Mellon University and was a Distinguished Member of the Technical Staff at Sandia National Laboratories. She has received numerous awards and honors, including election to the National Academy of Engineering in 2025, and has held leadership positions such as President of the American Institute of Mining, Metallurgical, and Petroleum Engineers in 2023.

Research topics

• Artificial Intelligence
• Computer Science
• Machine Learning
• Data science

Selected publications

• Combining multimodal fatigue fracture surface images for analysis with a CNN

  Scientific Reports · 2026-02-20 · 1 citations

  articleOpen access

  This work uses three different modalities, namely SEM, BSE and scanning white light interference (SWLI) to image fatigue fracture surfaces of Ti-6Al-4V. Convolutional neural networks (CNNs) that were pre-trained on images of the natural world were used to predict values such as distance from load line and crack growth rate. SEM images are routinely used to study the topography of fracture surfaces because the shallower interaction volume resolves surface features while BSE images and SWLI data add information about composition and surface height. Combining the three imaging modalities via the use of color channels facilitates overlaying them on the same grid for transferability of models pre-trained on colored images. This work shows that the imaging modalities under the guise of color channels have different levels of importance depending on the model being trained. It also documents how the combination of information from these modalities improves the classification and regression results by 20 and 60 %, respectively, relative to the secondary electron images alone. (162/200).

  PublisherDOI

• Evaluating and enhancing Segment Anything Model transferability for microstructural image analysis in nuclear materials

  Computational Materials Science · 2026-03-09

  articleOpen access

  Segmentation approaches in niche materials domains, such as in nuclear applications, can be hindered by limited data for neural network-based methodologies and inconsistent manual or algorithmic methods. In contrast, foundation models like Meta’s Segment Anything Model (SAM) offer potential zero-shot solutions. In this study, we evaluate SAM’s performance on three publicly available datasets relevant to nuclear materials research, spanning multiple microstructural features (Pt nanoparticles, dislocation loops with black dots, and cavities) that were imaged with varying electron microscopy modalities. Results from the SAM ViT-H model show a systematic tendency toward over-segmentation, which creates an opportunity for domain-informed segmentation filter development. Dataset-specific post-processing , including morphological filtering for complex features, enables segmentation results with F1 scores between 0.63 and 0.86, underscoring the role of domain knowledge in optimizing the segmentation transferability of this foundation model. This work identifies the opportunities and challenges of applying foundation models to microstructural image analysis in nuclear materials specifically and recommends the development of standardized filtering workflows to support broader community adoption. • SAM shows dataset-specific microstructural feature detection without fine-tuning. • SAM over-segmentation is corrected with post-processing and morphology filtering. • SAM transferability is assessed across electron microscopy imaging modalities and feature type. • SAM segmentation quality, with or without post-processing, affects property analysis.

  PublisherDOI

• Few-Shot Segmentation of Complex Microstructures: Minimizing Training Images by Maximizing Representativeness and Diversity

  SSRN Electronic Journal · 2026-01-01

  preprintOpen accessSenior author
  PublisherDOI

• On the Limits of Predictivity in Microstructural Evolution Simulations: Ensemble Simulations of Polycrystalline Grain Growth

  Metallurgical and Materials Transactions A · 2026-02-27

  articleOpen accessSenior author

  Abstract Theoretical models for polycrystalline grain growth are deterministic. However, although computer simulations based on these models reproduce average features and trends, they do not reliably predict experimental growth trajectories for individual grains. Disagreement between experiment and simulation has generally been attributed to shortcomings in the computational instantiation; however, even after several decades of improving the physical bases of computational models, a perfect match has not been achieved. In this study, we examine the sources of uncertainty in simulation and experiment during polycrystalline grain growth. In ensembles of nominally identical molecular dynamics simulations, growth trajectories of individual grains can vary significantly due to discrete events (topological transformations) that have cascading effects on the microstructural ensemble. The type and timing of these microstructural events are extraordinarily sensitive to atomic-scale processes. When we compare ensemble simulation results to experimental outcomes, we find that the ensemble simulation delineates a range of possible outcomes, and the experimental results fall within that range. The implication is that polycrystalline grain growth has a fundamental, aleatoric uncertainty that limits our ability to predict its outcomes. Simulation and experiment can never agree perfectly; however, simulations can predict the range of possible experimental outcomes.

  PublisherOA PDFDOI

• Mechanisms of shape evolution and rotation during crystal growth in cylindrical grains

  Acta Materialia · 2025-12-20

  articleOpen accessSenior authorCorresponding
  PublisherOA PDFDOI

• Mechanisms of Shape Evolution and Rotation during Crystal Growth in Cylindrical Grains

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Comparing molecular dynamics simulations of grain growth with experimental data

  Scripta Materialia · 2024-10-31 · 12 citations

  articleSenior authorCorresponding
  PublisherDOI

• Comparing Molecular Dynamics Simulations of Grain Growth with Experimental Data

  SSRN Electronic Journal · 2024-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Microscopy modality transfer of steel microstructures: Inferring scanning electron micrographs from optical microscopy using generative AI

  Materials Characterization · 2024-12-06 · 6 citations

  articleOpen accessSenior authorCorresponding
  PublisherOA PDFDOI

• Graph convolutional network for predicting abnormal grain growth in Monte Carlo simulations of microstructural evolution

  Scientific Reports · 2024-12-04 · 2 citations

  articleOpen accessSenior authorCorresponding

  Recent developments in graph neural networks show promise for predicting the occurrence of abnormal grain growth, which has been a particularly challenging area of research due to its apparent stochastic nature. In this study, we generate a large dataset of Monte Carlo simulations of abnormal grain growth. We train simple graph convolution networks to predict which initial microstructures will exhibit abnormal grain growth, and compare the results to a standard computer vision approach for the same task. The graph neural network outperformed the computer vision method and achieved 73% prediction accuracy and fewer false positives. It also provided some physical insight into feature importance and the relevant length scale required to maximize predictive performance. Analysis of the uncertainty in the Monte Carlo simulations provides additional insights for ongoing work in this area.

  PublisherOA PDFDOI

Recent grants

• Extracting Knowledge from 100 years of Microstructural Images: Using Machine Vision and Machine Learning to Address the Microstructural Big Data Challenge

  NSF · $400k · 2015–2018

• Coupled simulations of low temperature microstructural evolution in nanocrystalline metals

  NSF · $300k · 2013–2017

• CDS&E: A New Approach for Determining the Free Energy and Absolute Mobility of Flat, Curved, and Moving Interfaces

  NSF · $393k · 2018–2022

• QRM: Using Visual Information to Quantify Microstructure-Processing-Property Relationships

  NSF · $665k · 2018–2023

Frequent coauthors

• Carl Cady

  Los Alamos National Laboratory

  86 shared

• H. Henein

  University of Alberta

  86 shared

• Edward D. Herderick

  ASL Analytical (United States)

  86 shared

• Srinivas Chada

  Alcoa (United States)

  86 shared

• Angela C. Scott

  Mount Vernon Hospital

  72 shared

• Alexander P. Scott

  Centre for Environment, Fisheries and Aquaculture Science

  61 shared

• Ray D. Peterson

  61 shared

• G. W. Warren

  The University of Texas at Tyler

  50 shared

Education

• Dual PhD, Materials Science and Engineering and Scientific Computing

  University of Michigan

  1992

• S.M., Ceramics

  Massachusetts Institute of Technology

  1988

• B.S.E., Materials and Metallurgical Engineering

  University of Michigan

  1987

Awards & honors

• ASEE Mike Ashby Outstanding Materials Educator Award 2022
• AIME Honorary Member 2022
• ASM Edward DeMille Campbell Memorial Lecturer 2021
• TMS Alexander Scott Distinguished Service Award 2020
• Fellow of the Minerals, Metals, and Materials Society (FTMS)…