Materials AI and Scientific Language Models
Scientific language models for the literature on cement, inorganic glasses, and materials for extreme environments.

Materials for Extreme Environments, Cement, and Inorganic Glasses
Postdoctoral Researcher, Hopkins Extreme Materials Institute, Baltimore, Maryland, Johns Hopkins University
My research develops data-driven methods for materials in extreme environments and for construction materials, especially cement and inorganic glasses. I work on scientific language models, materials informatics, and explainable machine learning that connect composition, processing, testing, and properties. I am particularly interested in connecting scientific knowledge, experimental data, and computational models to accelerate materials discovery.
Scientific language models for the literature on cement, inorganic glasses, and materials for extreme environments.
Interpretable models for the composition, processing, and properties of cement and inorganic glasses.
Discovering and understanding materials that must perform in extreme environments.
Data-driven studies of cement, other construction materials, and inorganic glasses.
A family of large language models for materials research, with a study of model adaptability in continued pre-training.
Hopkins Extreme Materials Institute. Topic: discovering materials for extreme environments through agentic AI, with Prof. Michael D. Shields.
2026
@article{zaki2026llamat,
title = {LLaMat: A family of large language models for materials research with insights into model adaptability in continued pre-training},
author = {Zaki M, Ahlawat D, Mishra V, Singh S, Bihani V, Grover H, Mishra B, Miret S, Mausam M, Krishnan NMA},
year = {2026},
journal = {Nature Machine Intelligence},
}2024
@article{zaki2024mascqa,
title = {MaScQA: Investigating Materials Science Knowledge of Large Language Models},
author = {Zaki M, Jayadeva J, Mausam M, Anoop Krishnan NM},
year = {2024},
journal = {Digital Discovery},
doi = {10.1039/D3DD00188A},
}2023
@article{zaki2023glassomics,
title = {Glassomics: An omics approach toward understanding glasses through modeling, simulations, and artificial intelligence},
author = {Zaki M, Jan A, Krishnan NMA, Mauro JC},
year = {2023},
journal = {MRS Bulletin},
doi = {10.1557/s43577-023-00560-1},
}@article{zaki2023cementron,
title = {Cementron: Machine learning the alite and belite phases in cement clinker from optical images},
author = {Zaki M, Sharma S, Gurjar SK, Goyal R, Krishnan NA, others},
year = {2023},
journal = {Construction and Building Materials},
}@inproceedings{gupta2023discomat,
title = {DiSCoMaT: Distantly supervised composition extraction from tables in materials science articles},
author = {Gupta T, Zaki M, Khatsuriya D, Hira K, Krishnan NMA, Mausam M},
year = {2023},
booktitle = {Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
doi = {10.18653/v1/2023.acl-long.753},
}2022
@article{zaki2022afm-tip,
title = {Interpretable Machine Learning Approach for Identifying the Tip Sharpness in Atomic Force Microscopy},
author = {Zaki M, Kasimuthumaniyan S, Sahoo S, Jayadeva, Gosvami NN, Krishnan NMA},
year = {2022},
journal = {Scripta Materialia},
doi = {10.1016/j.scriptamat.2022.114965},
}@article{gupta2022matscibert,
title = {MatSciBERT: A materials domain language model for text mining and information extraction},
author = {Gupta T, Zaki M, Krishnan NMA, Mausam M},
year = {2022},
journal = {npj Computational Materials},
doi = {10.1038/s41524-022-00784-w},
}@article{zaki2022optical,
title = {Interpreting the optical properties of oxide glasses with machine learning and Shapely additive explanations},
author = {Zaki M, Venugopal V, Bhattoo R, et al.},
year = {2022},
journal = {Journal of the American Ceramic Society},
}2021
@article{venugopal2021looking,
title = {Looking through glass: Knowledge discovery from materials science literature using natural language processing},
author = {Venugopal V, Sahoo S, Zaki M, Agarwal M, Gosvami NN, Krishnan NMA},
year = {2021},
journal = {Patterns},
doi = {10.1016/j.patter.2021.100290},
}Selected studies that cite, extend, apply, or independently evaluate my research.
No verified records yet. A citation by itself is not treated as evidence that a study builds on this work.
Verified software and datasets will be listed here.
Reviewed more than 40 research papers.