Hugging Science

Community-driven open science for the age of AI

About

Many of the deepest challenges in advancing AI for scientific discovery are not purely technical—they are social and organizational. Progress is often limited not by algorithms or computational power, but by how effectively we coordinate efforts, share resources, and collaborate across disciplinary boundaries. Hugging Science brings together a global community of researchers, developers, and practitioners committed to accelerating breakthroughs in physics, biology, chemistry, neuroscience, and beyond through open collaboration.

Our vision is grounded in the argument presented in our position paper: democratizing AI for science requires treating it as a collective social project where equitable participation and sustainable collaboration are prerequisites for technical progress.

What We Do

• Launch collaborative challenges and open problem calls to identify and mobilize collective effort around upstream computational bottlenecks with broad applicability across scientific domains—such as efficient PDE solvers, multi-scale coupling, and high-dimensional sampling—rather than fragmenting resources across narrow, domain-specific applications

• Build open toolkits, benchmarks, and workflows that address data fragmentation through standardized formats and shared evaluation metrics, making it easier for researchers at institutions of all resource levels to collaborate and build on each other's work

• Support cross-disciplinary exchange and education by creating resources that bridge the communication gap between domain scientists who prioritize mechanistic understanding and ML researchers who focus on predictive performance, enabling more effective collaboration

• Nurture a community that values contributions to data curation, infrastructure development, and educational resources alongside algorithmic innovation—recognizing that datasets and infrastructure often have far greater long-term impact than individual models

• Learn together through open discussion of both technical advances and the social and institutional barriers that constrain progress, working to align incentives and build sustainable practices for scientific AI

Join Us

Discord: Join our server for real-time collaboration, discussion, and community coordination.

GitHub: Explore and contribute to our open-source projects and initiatives.

New to the community? Join the Discord to connect with other members and explore current initiatives. We'll help you find your place.

Have a project idea? You don't need permission to start building. If you'd like community support, computational resources, or collaborative connections, reach out via Discord or GitHub.

Learn More

Read our position paper: AI for Scientific Discovery is a Social Problem

Explore the Science Release Heatmap to visualize open models, datasets, and applications from organizations contributing to AI4Science, tracking release patterns across materials science, biology, physics, chemistry, and more!

Hugging Science is an open community effort. We believe that by working openly and collaboratively, we can make AI-enabled scientific discovery more accessible, equitable, and impactful.