Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 46 min ago
PreprintCC BYarXiv (Cornell University)

DPA4: Pushing the Accuracy-Cost Frontier of Interatomic Potentials with EMFA SO(2) Convolution

  • Tiancheng Li1,2
  • Wentao Li3
  • Anyang Peng1
  • Jianming Xue4,2
  • Linfeng Zhang1,5
  • Duo Zhang
  • Han Wang6,7
7 affiliations
  1. 1AI for Science Institute , Beijing 100080 , P. R. China
  2. 2State Key Laboratory of Nuclear Physics and Technology , School of Physics , Peking University , Beijing 100871 , China
  3. 3Department of Chemical Engineering , Tsinghua University , Beijing 100084 , P. R. China
  4. 4Center for Applied Physics and Technology , Peking University , Beijing 100871 , China
  5. 5DP Technology , Beijing 100080 , P. R. China
  6. 6HEDPS, CAPT , College of Engineering , Peking University , Beijing 100871 , P. R. China
  7. 7National Key Laboratory of Computational Physics , Institute of Applied Physics and Computational Mathematics , Fenghao East Road 2 , Beijing 100094 , P. R. China
Open at sourcePDFDOICite

Abstract

Machine-learning interatomic potentials now approach quantum-mechanical accuracy, but the most expressive equivariant architectures are costly to evaluate, and the leading ones depend on auxiliary denoising or direct-force pretraining. We introduce DPA4, an SE(3)-equivariant architecture spanning six size classes from 0.48 to 25 million parameters and reaching the accuracy of the strongest published models at several-fold to an order-of-magnitude higher inference throughput. Its convolution couples edge and node features across all angular degrees in an edge-local frame, and its Wigner bilinear nonlinearity is universal in its full form and, on an exact quadrature grid, equivariant to machine precision. On Matbench Discovery, DPA4 leads every ranked metric, and every variant evaluated lies on the accuracy--throughput Pareto frontier of the compliant leaderboard. DPA4-Pro attains the lowest energy error on OMat24 and lower total-energy and force errors than the strongest conservative baseline on the OMol25 composition-validation split. All variants are trained through the conservative energy-gradient path alone, made practical by a threefold-faster compiled implementation. DPA4 thus brings leaderboard-class accuracy within the routine compute budgets of molecular-dynamics and materials-screening workflows, for both inorganic crystals and organic molecules.

Related

  1. Influence of exchange–correlation functional on machine-learnt interatomic potentials accuracy: a systematic study in borosilicate glasses
    Fengming Shi, Luca Brugnoli, François-Xavier Coudert
    ChemRxivPreprintCC BYMaterials & polymersChemRxiv
  2. Influence of Exchange–Correlation Functional on Machine-Learned Interatomic Potentials’ Accuracy: A Systematic Study of Borosilicate Glasses
    Fengming Shi, Luca Brugnoli, François-Xavier Coudert
    Journal of Chemical Theory and ComputationPeer-reviewedCC BY 4.0Full textMaterials & polymersTheoretical & computational chemistryHAL
  3. Pushing the accuracy of on-top functionals with agent-driven supervised learning
    Yuhao Chen, Dayou Zhang, Donald G. Truhlar +1
    arXivPreprintarXiv licenceFull textPhysical chemistryarXiv
  4. Correcting the Liquid Densities of Universal Machine-Learning Interatomic Potentials Using Local-Number-Density Repulsion
    Shunsuke Mieda
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv
  5. Foundation interatomic potentials learn the practical chemistry they were never taught
    Alexander A. Kolganov, Mas P. Klein, Evgeny A. Pidko
    ChemRxivPreprintLicence not statedGeneral chemistryChemRxiv

Cite

BibTeXRIS
BibTeX
@misc{li2026dpa4,
  title = {{DPA4: Pushing the Accuracy-Cost Frontier of Interatomic Potentials with EMFA SO(2) Convolution}},
  author = {Tiancheng Li and Wentao Li and Anyang Peng and Jianming Xue and Linfeng Zhang and Duo Zhang and Han Wang},
  howpublished = {arXiv (Cornell University)},
  year = {2026},
  month = jun,
  doi = {10.48550/arxiv.2606.02419},
  url = {https://arxiv.org/abs/2606.02419},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - DPA4: Pushing the Accuracy-Cost Frontier of Interatomic Potentials with EMFA SO(2) Convolution
AU  - Tiancheng Li
AU  - Wentao Li
AU  - Anyang Peng
AU  - Jianming Xue
AU  - Linfeng Zhang
AU  - Duo Zhang
AU  - Han Wang
PY  - 2026
DA  - 2026/06/01
PB  - arXiv (Cornell University)
DO  - 10.48550/arxiv.2606.02419
UR  - https://arxiv.org/abs/2606.02419
AB  - Machine-learning interatomic potentials now approach quantum-mechanical accuracy, but the most expressive equivariant architectures are costly to evaluate, and the leading ones depend on auxiliary denoising or direct-force pretraining. We introduce DPA4, an SE(3)-equivariant architecture spanning six size classes from 0.48 to 25 million parameters and reaching the accuracy of the strongest published models at several-fold to an order-of-magnitude higher inference throughput. Its convolution couples edge and node features across all angular degrees in an edge-local frame, and its Wigner bilinear nonlinearity is universal in its full form and, on an exact quadrature grid, equivariant to machine precision. On Matbench Discovery, DPA4 leads every ranked metric, and every variant evaluated lies on the accuracy--throughput Pareto frontier of the compliant leaderboard. DPA4-Pro attains the lowest energy error on OMat24 and lower total-energy and force errors than the strongest conservative baseline on the OMol25 composition-validation split. All variants are trained through the conservative energy-gradient path alone, made practical by a threefold-faster compiled implementation. DPA4 thus brings leaderboard-class accuracy within the routine compute budgets of molecular-dynamics and materials-screening workflows, for both inorganic crystals and organic molecules.
SN  - 2331-8422
ER  -