We encourage participation on topics that explore any form of synergy between scientific principles and artificial intelligence (AI)/ML methods. Examples of relevant submissions include (but are not limited to):
AI/ML algorithms that employ soft or hard scientific constraints in the learning process.
Physics-informed neural networks for solving partial differential equations.
Modeling multi-scale multi-physics phenomena.
Methods to encode scientific knowledge in AI/ML model architecture.
Science guided generative or reinforcement learning methods.
Approaches that use scientific knowledge for interpreting ML results along the lines of explainable AI.
Surrogate and reduced order modeling methods.
Parameterization and downscaling methods.
AI/ML methods for discovering governing equations from data.
Hybrid constructions of science-based & AI/ML-based models.
Software development facilitating the inclusion of scientific knowledge in learning
Inverse modeling & system identification techniques using AI.
Techniques for using data to calibrate parameters and system states in scientific models.
We are soliciting paper submissions for position, review, or research articles in two formats: (i) short papers (2-4 pages, excluding references) and (ii) full papers (6-8 pages, excluding references). Extended versions of articles in submission at other venues are acceptable as long as they do not violate the dual-submission policy of the other venue. We also encourage early drafts of on-going research with preliminary insights/results that contribute to the symposium agenda. All submissions will undergo peer review and authors will have the option to publish their work in an open access proceedings site.
Submissions should be formatted according to the AAAI template (see Author Kit) and submitted via EasyChair. All submissions will undergo double-blind peer review before acceptance.