EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Generative pseudo-force fields for molecular generation

Stefaan Simon Pierre Hessmann, Khaled Kahouli, Stefan Gugler, Michael Plainer, Frank Noé, Klaus-Robert Mueller, Niklas Wolf Andreas Gebauer

Machine Learning: Science and Technology · 16.11.2026

DOI: 10.1088/2632-2153/aec2a2 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning: Science and Technology, online erschienen am 16.11.2026. Verfasst von Stefaan Simon Pierre Hessmann und 6 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Institute of Information & communications Technology Planning & Evaluation, Korea government, Postdoc.Mobility.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Generating stable molecular conformations typically forces a tradeoff between the physical realism of energy-based relaxation and the sampling efficiency of data-driven generative models. While machine learning force fields (MLFFs) can sample stable conformations by relaxing molecular geometries according to physical forces, they require costly ab-initio training data. Conversely, diffusion models (DMs) learn from equilibrium data alone but are dependent on noise schedules and time-step conditioning. In this work, we propose generative pseudo-force fields (GPFFs) to bridge these paradigms by training an MLFF on a quadratic pseudo-potential energy surface relative to reference equilibrium structures. Because no ab-initio calculations are required for the perturbed geometries, non-equilibrium training data can be generated on the fly by perturbing the equilibria with Gaussian noise. We show that GPFFs constitute a time-step-agnostic variant of variance-exploding DMs: the score comes from the predicted pseudo-forces but because force magnitudes implicitly encode the noise level, no time-step conditioning is needed. Our GPFF can hence be used as a drop-in replacement in standard diffusion sampling (ancestral, Heun) but also facilitates more efficient, adaptive variants and an MLFF inspired direct denoising scheme. Our proposed sampling algorithms support arbitrary structural priors and geometric constraints. On QM9, GPFF reaches 99 % validity with stochastic direct denoising at 256 neural function evaluations (NFE), and about 90 % at just 16 NFE with direct denoising, outperforming a matched
diffusion baseline that shares its backbone, training recipe, and noise schedule across all samplers. Combined with custom priors, we showcase the fast and accurate generation process of our method in a molecular editor for a drug design setting.

Förderung

  • Institute of Information & communications Technology Planning & Evaluation
  • Korea government
  • Postdoc.Mobility
  • Berlin Mathematics center MATH+
  • of Korea University
  • Deutsche Forschungsgemeinschaft
  • Federal Ministry of Education and Research
  • German Federal Ministry for Research, Technology and Space

Zitieren

APA

Hessmann, S. S. P., Kahouli, K., Gugler, S., Plainer, M., Noé, F., Mueller, K., & Gebauer, N. W. A. (2026). Generative pseudo-force fields for molecular generation. Machine Learning: Science and Technology. https://doi.org/10.1088/2632-2153/aec2a2

Vancouver

Hessmann SSP, Kahouli K, Gugler S, Plainer M, Noé F, Mueller K, et al. Generative pseudo-force fields for molecular generation. Machine Learning: Science and Technology. 2026. doi:10.1088/2632-2153/aec2a2

BibTeX
@article{Hessmann2026,
  author = {Hessmann, Stefaan Simon Pierre and Kahouli, Khaled and Gugler, Stefan and Plainer, Michael and Noé, Frank and Mueller, Klaus-Robert and Gebauer, Niklas Wolf Andreas},
  title = {Generative pseudo-force fields for molecular generation},
  journal = {Machine Learning: Science and Technology},
  year = {2026},
  doi = {10.1088/2632-2153/aec2a2}
}
Kostenlos starten →