Oct
15
Open community · Est. 2026
Where machine learning meets the laws of nature.
AI4Physics brings together physicists and AI researchers to build learning systems that respect physical law — and to use them to discover new physics. Seminars, workshops, open resources, and a place to collaborate.
500+Members worldwide
40+Seminar talks
6Research themes
2×Workshops per year
About
A shared language for two fields.
Physics has centuries of hard-won structure — symmetries, conservation laws, effective theories. AI brings powerful tools for learning from data at scale. The most exciting progress happens when the two talk to each other.
We are an open, non-commercial community hosted by researchers across universities and labs. Everyone is welcome — from undergraduates to senior faculty, from theorists to experimentalists to ML engineers.
01
Connect
Monthly online seminars and an active community forum.
02
Learn
Curated tutorials, datasets, benchmarks, and reading lists.
03
Build
Open-source projects, hackathons, and cross-disciplinary collaborations.
Research themes
What we work on
Six threads that run through our talks, workshops, and projects.
Neural PDE solvers & surrogates
Neural operators, PINNs, and learned simulators that accelerate fluid, plasma, and climate models by orders of magnitude.
Symmetry-aware learning
Equivariant architectures and geometric deep learning that bake in conservation laws and invariances.
Scientific discovery
Symbolic regression, automated hypothesis generation, and AI agents that propose and test physical models.
AI for experiments
Real-time triggers, anomaly detection, and autonomous control for colliders, telescopes, and quantum labs.
Foundation models for physics
Large pretrained models across physical systems, multimodal field data, and scientific reasoning benchmarks.
Physics of learning
Statistical mechanics, scaling laws, and renormalization-group views of how and why deep networks work.
Upcoming
Events
Seminars are online and free. Workshops rotate between host institutions.
Seminar
Nov
05
Tutorial: Equivariant networks from scratch
Dec
03
Seminar: LLM agents as research collaborators
Mar
'27
AI4Physics Spring Workshop 2027
Resources
Start here
Curated by the community. Suggestions welcome.
READ
LEARN
DATA
CODE
Reading list
Foundational and recent papers, organized by theme and difficulty.
Tutorials & notebooks
Hands-on Colab notebooks: PINNs, neural operators, equivariant GNNs.
Datasets & benchmarks
A living index of open physics datasets and ML benchmarks.
Open-source projects
Community-maintained code on GitHub. Contributions welcome.
Organizers
The team
A
Organizer Name
Institution · Role
B
Organizer Name
Institution · Role
C
Organizer Name
Institution · Role
D
Organizer Name
Institution · Role
Join us
Be part of AI4Physics.
Get seminar announcements, workshop calls, and a monthly digest of the best new work at the intersection of AI and physics.
No spam. Unsubscribe anytime.