Xiaoting Zhong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xiaoting Zhong has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
Brian Gallagher
Center for Applied Scientific Computing, Lawrence Livermore National Laboratory, Livermore, USA
2 shared publications- Thomas Yong-Jin Han
Materials Science Division, Lawrence Livermore National Laboratory, Livermore, CA, 94550, USA
2 shared publications - Anna M. Hiszpanski
Lawrence Livermore National Laboratory, Livermore, USA
1 shared publication Bhavya Kailkhura
Center for Applied Scientific Computing, Lawrence Livermore National Laboratory, Livermore, USA
1 shared publicationEmily Robertson
Materials Science Division, Lawrence Livermore National Laboratory, Livermore, USA
1 shared publication- Gregory S. Rohrer
Department of Materials Science and Engineering Carnegie Mellon University Pittsburgh Pennsylvania USA
1 shared publication
Latest Publications
- Top Cited PapersOpen Access
- 22 September 2022
- Published by Springer Nature in npj Computational Materials
- Vol. 8 (1) , 1-19
Explainable machine learning in materials science
Machine learning models are increasingly used in materials studies because of their exceptional accuracy. However, the most accurate machine learning models are usually difficult to explain. Remedies to this problem lie in explainable artificial intelligence (XAI), an emerging research field that addresses the explainability of complicated machine learning models like deep neural networks (DNNs). This article attempts to provide an entry point to XAI for materials scientists. Concepts are defined to clarify what explain means in the context of materials science. Example works are reviewed to show how XAI helps materials science research. Challenges and opportunities are also discussed.
- 22 September 2022
- Published by Springer Nature in npj Computational Materials
- Vol. 8 (1) , 1-19
Top Cited PapersOpen AccessExplainable machine learning in materials science
Machine learning models are increasingly used in materials studies because of their exceptional accuracy. However, the most accurate machine learning models are usually difficult to explain. Remedies to this problem lie in explainable artificial intelligence (XAI), an emerging research field that addresses the explainability of complicated machine learning models like deep neural networks (DNNs). This article attempts to provide an entry point to XAI for materials scientists. Concepts are defined to clarify what explain means in the context of materials science. Example works are reviewed to show how XAI helps materials science research. Challenges and opportunities are also discussed.
- Open Access
- 5 October 2021
- Published by Springer Nature in npj Computational Materials
- Vol. 7 (1) , 1-11
A study of real-world micrograph data quality and machine learning model robustness
Machine-learning (ML) techniques hold the potential of enabling efficient quantitative micrograph analysis, but the robustness of ML models with respect to real-world micrograph quality variations has not been carefully evaluated. We collected thousands of scanning electron microscopy (SEM) micrographs for molecular solid materials, in which image pixel intensities vary due to both the microstructure content and microscope instrument conditions. We then built ML models to predict the ultimate compressive strength (UCS) of consolidated molecular solids, by encoding micrographs with different image feature descriptors and training a random forest regressor, and by training an end-to-end deep-learning (DL) model. Results show that instrument-induced pixel intensity signals can affect ML model predictions in a consistently negative way. As a remedy, we explored intensity normalization techniques. It is seen that intensity normalization helps to improve micrograph data quality and ML model robustness, but microscope-induced intensity variations can be difficult to eliminate.
- 5 October 2021
- Published by Springer Nature in npj Computational Materials
- Vol. 7 (1) , 1-11
Open AccessA study of real-world micrograph data quality and machine learning model robustness
Machine-learning (ML) techniques hold the potential of enabling efficient quantitative micrograph analysis, but the robustness of ML models with respect to real-world micrograph quality variations has not been carefully evaluated. We collected thousands of scanning electron microscopy (SEM) micrographs for molecular solid materials, in which image pixel intensities vary due to both the microstructure content and microscope instrument conditions. We then built ML models to predict the ultimate compressive strength (UCS) of consolidated molecular solids, by encoding micrographs with different image feature descriptors and training a random forest regressor, and by training an end-to-end deep-learning (DL) model. Results show that instrument-induced pixel intensity signals can affect ML model predictions in a consistently negative way. As a remedy, we explored intensity normalization techniques. It is seen that intensity normalization helps to improve micrograph data quality and ML model robustness, but microscope-induced intensity variations can be difficult to eliminate.
- 24 May 2019
- Published by Wiley in Journal of the American Ceramic Society
- Vol. 102 (11) , 7003-7014
Grain boundary curvatures in polycrystalline SrTiO3: Dependence on grain size, topology, and crystallography
By mapping grain orientations on parallel serial sections of a SrTiO3 ceramic, it was possible to reconstruct three‐dimensional orientation maps containing more than 3000 grains. The grain boundaries were approximated by a continuous mesh of triangles and mean curvatures were determined for each triangle. The integral mean curvatures of grain faces were determined for all grains. Small grains with fewer than 16 neighbors mostly have positive mean curvatures while larger grains with more than 16 neighbors mostly have negative mean curvatures. It is also possible to correlate the mean curvature of individual triangles with the crystallographic characteristics of the grain boundary. The mean curvature is lowest for grain boundaries with (100) orientations and highest for grain boundaries with (111) orientations. This trend is inversely correlated to the relative areas of grain boundaries and directly correlated to the relative grain boundary energy. The direct correlation between the energy and curvature is consistent with the expected behavior of grain boundaries made up of singular orientations. Furthermore, because both the relative energy and curvature of grain boundaries with (100) orientations are minima in the distributions, these boundaries also have the lowest driving force for migration.
- 24 May 2019
- Published by Wiley in Journal of the American Ceramic Society
- Vol. 102 (11) , 7003-7014
Grain boundary curvatures in polycrystalline SrTiO3: Dependence on grain size, topology, and crystallography
By mapping grain orientations on parallel serial sections of a SrTiO3 ceramic, it was possible to reconstruct three‐dimensional orientation maps containing more than 3000 grains. The grain boundaries were approximated by a continuous mesh of triangles and mean curvatures were determined for each triangle. The integral mean curvatures of grain faces were determined for all grains. Small grains with fewer than 16 neighbors mostly have positive mean curvatures while larger grains with more than 16 neighbors mostly have negative mean curvatures. It is also possible to correlate the mean curvature of individual triangles with the crystallographic characteristics of the grain boundary. The mean curvature is lowest for grain boundaries with (100) orientations and highest for grain boundaries with (111) orientations. This trend is inversely correlated to the relative areas of grain boundaries and directly correlated to the relative grain boundary energy. The direct correlation between the energy and curvature is consistent with the expected behavior of grain boundaries made up of singular orientations. Furthermore, because both the relative energy and curvature of grain boundaries with (100) orientations are minima in the distributions, these boundaries also have the lowest driving force for migration.