Hugging Face Introduces DOIs for Datasets and Models

Reproducibility and citability are pillars of scientific research. Until now, citing a dataset or AI model hosted on Hugging Face could be complex due to the lack of a persistent identification system. To address this need, Hugging Face has announced the introduction of DOIs (Digital Object Identifiers) for its resources.\n\n## What is a DOI and why is it crucial for AI?\nA Digital Object Identifier (DOI) is a unique and persistent identifier used to reliably cite digital objects, such as research papers, data, or software. By integrating DOIs, Hugging Face allows researchers and developers to assign a precise reference to their work, facilitating discovery, access, and citation in academic publications.\n\nThis initiative is a major step toward recognizing the work done by contributors in the Open Source and scientific communities: every version of a model or dataset can now be identified without ambiguity.\n\n## Benefits for the community\nThe introduction of DOIs on the platform offers three main advantages:\n- Easier citations: Researchers can now include direct and stable links in their bibliographies, ensuring that readers access the exact version of the model or dataset used.\n- Credibility and recognition: By making work as citable as any other scientific object, this measure strengthens the legitimacy of AI models as major academic contributions.\n- Increased reproducibility: Version management associated with DOIs ensures that a researcher working on an experiment can find the exact state of the resources used by another, even years later.\n\n## How to get a DOI on Hugging Face?\nThe implementation is designed to be as seamless as possible for creators. You can now simply submit a request via the object management interface (dataset or model). Once validated, an official DOI is generated and displayed on the resource page. This DOI points to a persistent page, ensuring access to the resource even if the internal Hugging Face URL were to change.\n\nThis advancement is in line with the platform's ongoing efforts to structure the open AI ecosystem. By providing these information management tools, Hugging Face is no longer just a repository, but is establishing itself as essential infrastructure for research in generative artificial intelligence and machine learning.\n\nFor research teams, this ensures that their investments in time and resources to train models or clean data will be properly attributed by their peers, thereby strengthening the culture of sharing and openness within data science.