Call for Papers

We invite paper submissions to the workshop on Foundation Models for Science (FM4Science) co-located with the ICLR 2026, the Fourteenth International Conference on Learning Representations, at Rio de Janeiro, Brazil.

Topics will include, but are not limited to:

  • How to achieve better scaling laws for foundation models in scientific problems by designing datasets, network architectures, and training algorithms?
  • How to design data augmentation and multi-modal self-supervised pretraining for scientific problems?
  • How to design efficient fine-tuning with scientific awareness?
  • How to quantify and reduce the uncertainty of scientific foundation models?
  • How to improve the out-of-distribution generalization of scientific foundation models?
  • How to make foundation models compatible with and enable the integration of classic scientific tools (simulators, solvers, etc.)?
  • How to make scientific foundation models benefit from the reasoning and in-context learning of LLMs?
  • How to better combine symbolic learning and data-driven learning?
  • How to use foundation models to facilitate visualizations in scientific problems?
  • How to accelerate scientific discovery and the collection/assimilation of scientific data with foundation models?
  • How to diagnose failure cases or modes where scientific foundation models do not perform well?
  • How to align scientific foundation models with scientific facts without hallucination?

Scientific Domains. We invite paper submissions from various scientific domains, including but not limited to: Astrophysics and Space Science, Biomedicine (e.g., proteins, biosequences, virtual screening), Computational Science (e.g., PDEs, forecasting), Earth Science, Materials Science (e.g., batteries, chemical synthesis), Quantum Mechanics (e.g., nuclear fusion), Small Molecules. Applications-driven submissions focusing on AI-for-Science and Scientific Machine Learning (SciML) are also highly encouraged.

Tentative important dates (AoE time):

Awards

Among exceptional research papers with high review scores, we will select one best paper award and two runner-ups.

Submission policy

Submission Format (inherit from ICLR 2026): Latex Template.

Submission portal: OpenReview.

Our submission policy is inherited from ICLR 2026. All authors and submissions must adhere to the ICLR code of conduct.

Contact

If you have any questions about paper submission and the workshop, please send email to: foundationmodelscience@gmail.com.