Universal MLIPs | Matthias Kellner

I am actively involved in the development of universal machine learning interatomic potentials (uMLIPs).

I am a co-author of the PET-MAD model (Mazitov et al., 2025), a uMLIP that emphasizes data quality over quantity.

I have developed the PET-MOLS universal organic force field, a transferable machine learning potential to study organic solids and amorphous organic solids, and have applied it to amorphously formulated active pharmaceutical ingredients (Kellner et al., 2026).

This work builds on the metatensor and metatomic libraries for interoperable atomistic machine learning (Bigi et al., 2026) and on the i-PI simulation engine (Litman et al., 2024), both of which I contribute to.

References

2026

arXiv

Quantum-corrected NMR crystallography at scale

Matthias Kellner, Ruben Rodriguez-Madrid, Jacob B Holmes, and 3 more authors

arXiv preprint, 2026

arXiv Bib

@article{kellner2026quantum,
  title = {Quantum-corrected NMR crystallography at scale},
  author = {Kellner, Matthias and Rodriguez-Madrid, Ruben and Holmes, Jacob B and Principe, Victor Paul and Emsley, Lyndon and Ceriotti, Michele},
  journal = {arXiv preprint},
  year = {2026},
}

JCP

Metatensor and metatomic: foundational libraries for interoperable atomistic machine learning

Filippo Bigi, Joseph W Abbott, Philip Loche, and 12 more authors

The Journal of Chemical Physics, 2026

DOI Bib Code

@article{bigi2026metatensor,
  title = {Metatensor and metatomic: foundational libraries for interoperable atomistic machine learning},
  author = {Bigi, Filippo and Abbott, Joseph W and Loche, Philip and Mazitov, Arslan and Tisi, Davide and Langer, Marcel F and Goscinski, Alexander and Pegolo, Paolo and Chong, Sanggyu and Goswami, Rohit and Febrer, Pol and Chorna, Sofiia and Kellner, Matthias and Ceriotti, Michele and Fraux, Guillaume},
  journal = {The Journal of Chemical Physics},
  volume = {164},
  number = {6},
  pages = {064113},
  year = {2026},
  publisher = {AIP Publishing},
  doi = {10.1063/5.0304911},
}

2025

Nat. Commun.

PET-MAD as a lightweight universal interatomic potential for advanced materials modeling

Arslan Mazitov, Filippo Bigi, Matthias Kellner, and 6 more authors

Nature Communications, 2025

DOI Bib Code

@article{mazitov2025pet,
  title = {PET-MAD as a lightweight universal interatomic potential for advanced materials modeling},
  author = {Mazitov, Arslan and Bigi, Filippo and Kellner, Matthias and Pegolo, Paolo and Tisi, Davide and Fraux, Guillaume and Pozdnyakov, Sergey and Loche, Philip and Ceriotti, Michele},
  journal = {Nature Communications},
  volume = {16},
  number = {1},
  pages = {10653},
  year = {2025},
  publisher = {Nature Publishing Group},
  doi = {10.1038/s41467-025-65662-7},
}

2024

JCP

i-PI 3.0: A flexible and efficient framework for advanced atomistic simulations

Yair Litman, Venkat Kapil, Yotam MY Feldman, and 13 more authors

The Journal of Chemical Physics, 2024

DOI Bib Code

@article{litman2024pi,
  title = {i-PI 3.0: A flexible and efficient framework for advanced atomistic simulations},
  author = {Litman, Yair and Kapil, Venkat and Feldman, Yotam MY and Tisi, Davide and Begu{\v{s}}i{\'c}, Tomislav and Fidanyan, Karen and Fraux, Guillaume and Higer, Jacob and Kellner, Matthias and Li, Tao E and P{\'o}s, Eszter S and Stocco, Elia and Trenins, George and Hirshberg, Barak and Rossi, Mariana and Ceriotti, Michele},
  journal = {The Journal of Chemical Physics},
  volume = {161},
  number = {6},
  pages = {062504},
  year = {2024},
  publisher = {AIP Publishing},
  doi = {10.1063/5.0215869},
}