Call for Papers

We invite submissions to the ICLR 2023 workshop on Physics for Machine Learning. The total length of the initial (and final) submission must not exceed 4 pages, with unlimited additional pages for citations. Authors may include an appendix after the references. However, reviewers are not required to read the appendix. Submissions should be a single file in .pdf format using this style file. The review process is double-blind, so please ensure that all papers are appropriately anonymized. We reserve the right to desk-reject improperly-anonymized papers.

This workshop is non-archival; even though all accepted papers will be available on OpenReview, there are no formally-published proceedings.

Submission link	OpenReview
Submission opens	1 January 2023
Submission deadline	8 February 2023 (AoE)
Final decisions	3 March 2023

List of topics

We invite all submissions on using physics for machine learning methods. A list of exemplary topics can be found below. Please note that this list is non-exhaustive. If you are not sure if your topic is suitable for the workshop, please feel free to contact any of the organizers.

• Physics-inspired machine learning; in particular for
  • Graph representation learning
  • Sequence modeling (e.g. Transformers, RNNs)
  • Generative modeling (e.g. diffusion models, score-based SDEs, normalizing flows)
  • Neural ODEs (e.g. NCDEs, CNFs)
• Equivariant neural networks
• Physics-based optimization
• Machine learning methods with a physics-based inductive bias, for instance applied to
  • Molecular simulations
  • Fluid dynamics
  • Astrophysics
  • Particle physics
  • Multi-scale problems (e.g. in multi-physics)
• Physics-based symbolic regression
• Dynamical systems reconstruction with physics-based inductive bias