Spaces:
Running
Running
| <html> | |
| <head> | |
| <meta charset="utf-8"> | |
| <meta name="description" | |
| content="PARTAGES"> | |
| <meta name="keywords" content="Generative AI, Health, Medical report"> | |
| <meta name="viewport" content="width=device-width, initial-scale=1"> | |
| <title>PARTAGES</title> | |
| <!-- CSS --> | |
| <link rel="stylesheet" href="./static/css/bulma.min.css"> | |
| <link rel="stylesheet" href="./static/css/bulma-carousel.min.css"> | |
| <link rel="stylesheet" href="./static/css/bulma-slider.min.css"> | |
| <link rel="stylesheet" href="./static/css/fontawesome.all.min.css"> | |
| <!-- Academicons (pour icône arXiv etc.) --> | |
| <link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css"> | |
| <!-- Fonts --> | |
| <link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"> | |
| <!-- Ton style perso --> | |
| <link rel="stylesheet" href="./static/css/index.css"> | |
| <script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script> | |
| <script defer src="./static/js/fontawesome.all.min.js"></script> | |
| <script src="./static/js/bulma-carousel.min.js"></script> | |
| <script src="./static/js/bulma-slider.min.js"></script> | |
| <script src="./static/js/index.js"></script> | |
| </head> | |
| <body> | |
| <section class="hero"> | |
| <div class="hero-body"> | |
| <div class="container is-max-desktop"> | |
| <div class="columns is-centered"> | |
| <div class="column has-text-centered"> | |
| <h1 class="title is-1 publication-title">PARTAGES</h1> | |
| <div class="is-size-5 publication-authors"> | |
| <span class="author-block"> | |
| <a href="https://www.health-data-hub.fr/" target="_blank">Health Data Hub</a><sup>1</sup>,</span> | |
| <span class="author-block"> | |
| <a href="https://www.lisn.upsaclay.fr/" target="_blank">LISN</a><sup>2</sup>,</span> | |
| <span class="author-block"> | |
| <a href="https://www.greyc.fr/" target="_blank">GREYC</a><sup>2</sup>, | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.liglab.fr/fr" target="_blank">LIG</a><sup>2</sup>, | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.limics.fr/" target="_blank">LIMICS</a><sup>2</sup>, | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://lia.univ-avignon.fr/" target="_blank">LIA</a><sup>1,2</sup>, | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.ls2n.fr/" target="_blank">LS2N</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.lis-lab.fr/" target="_blank">LIS</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.uness.fr/" target="_blank">UNESS</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.bordeaux-population-health.center/fr/" target="_blank">Bordeaux Population Health</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.cnrs.fr/fr" target="_blank">CNRS</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.inria.fr/fr" target="_blank">INRIA</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.aphp.fr/" target="_blank">AP-HP</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.centreleonberard.fr/" target="_blank">Centre Léon Bérard</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-rouen.fr/" target="_blank">CHU de Rouen</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://teamhcl.chu-lyon.fr/groupement-de-cooperation-sanitaire-houraa-gcs-houraa" target="_blank">GCS HOURAA</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.gustaveroussy.fr/" target="_blank">Institut Gustave Roussy</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://curie.fr/" target="_blank">Institut Curie</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.hpsj.fr/" target="_blank">Hôpitaux Saint-Joseph & | |
| Marie-Lannelongue</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.hopital-foch.com/" target="_blank">Hôpital Foch</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.elsan.care/fr" target="_blank">ELSAN</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="google.com/search?q=CHU+de+Toulouse&rlz=1C1GCEA_enFR1064FR1064&oq=CHU+de+Toulouse&gs_lcrp=EgZjaHJvbWUyCQgAEEUYORiABDIHCAEQABiABDIHCAIQABiABDIHCAMQABiABDIHCAQQABiABDIHCAUQABiABDIHCAYQABiABDIHCAcQABiABDIHCAgQABiABDIHCAkQABiABNIBBzgzMWowajSoAgCwAgA&sourceid=chrome&ie=UTF-8" target="_blank">CHU de Toulouse</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-bordeaux.fr/" target="_blank">CHU de Bordeaux</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-amiens.fr/" target="_blank">CHU Amiens-Picardi</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-brest.fr/" target="_blank">CHU Brest</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-lille.fr/" target="_blank">CHU de Lille</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-nancy.fr/" target="_blank">CHRU de Nancy</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.chu-reunion.fr/" target="_blank">CHU de La Réunion</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://www.ramsaysante.fr/" target="_blank">Ramsay Santé</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://recital.ai/" target="_blank">reciTAL</a><sup>2</sup> | |
| </span> | |
| <span class="author-block"> | |
| <a href="https://mistral.ai/fr" target="_blank">Mistral AI</a><sup>2</sup> | |
| </span> | |
| </div> | |
| <!-- Data --> | |
| <div class="column has-text-centered"> | |
| <div class="publication-links"> | |
| <!-- Code Link. --> | |
| <span class="link-block"> | |
| <a href="https://huggingface.co/HealthDataHub/models" target="_blank" | |
| class="external-link button is-normal is-rounded is-dark"> | |
| <span class="icon"> | |
| <i class="fab fa-github"></i> | |
| </span> | |
| <span>Code</span> | |
| </a> | |
| </span> | |
| <!-- Dataset Link. --> | |
| <span class="link-block"> | |
| <a href="https://huggingface.co/HealthDataHub/datasets" target="_blank" | |
| class="external-link button is-normal is-rounded is-dark"> | |
| <span class="icon"> | |
| <i class="far fa-images"></i> | |
| </span> | |
| <span>Data</span> | |
| </a> | |
| </span> | |
| </div> | |
| </div> | |
| </div> | |
| </div> | |
| </div> | |
| </div> | |
| </section> | |
| <section class="section"> | |
| <div class="container is-max-desktop"> | |
| <!-- Logo --> | |
| <div class="has-text-centered" style="margin-bottom: 2rem;"> | |
| <img src="./static/images/PARTAGES BASELINE_RVB.png" alt="Project logo" style="max-height: 120px;"> | |
| </div> | |
| <!--/ Logo --> | |
| <!-- Abstract. --> | |
| <div class="columns is-centered has-text-centered"> | |
| <div class="column is-four-fifths"> | |
| <h2 class="title is-3">Abstract</h2> | |
| <div class="content has-text-justified"> | |
| <p> | |
| <strong>PARTAGES (Advanced Development of Digital Commons for Generative Artificial Intelligence in Healthcare)</strong> is a project coordinated by the Health Data Hub. A winner of the “Digital Commons for Generative AI” call for projects under the France 2030 plan, it aims to accelerate and democratize the use of large language models (LLMs) for healthcare professionals. | |
| </p> | |
| <p> | |
| <strong>Its goal: </strong> to create a national momentum fostering the emergence of open generative AI solutions in healthcare, as well as their adoption within the healthcare ecosystem—whether academic, research-based, or industrial. | |
| </p> | |
| </div> | |
| </div> | |
| </div> | |
| <!--/ Abstract. --> | |
| <!-- Text. --> | |
| <div class="columns is-centered"> | |
| <div class="column is-four-fifths"> | |
| <h2 class="title is-3">A major national project</h2> | |
| <div class="content has-text-justified"> | |
| <p> | |
| With a budget of <strong>€9.4 million</strong>, PARTAGES is supported by a <strong>unique consortium of 32 partners</strong> mobilized on a national scale: | |
| </p> | |
| <ul> | |
| <li><strong>10 research teams</strong> (CNRS, Inria, universities),</li> | |
| <li><strong>20 public and private healthcare institutions</strong> (AP-HP, Institut Curie, Centre Léon Bérard, Ramsay Santé, ELSAN, 12 university hospitals...),</li> | |
| <li><strong>DeepTech companies specializing in AI</strong>, including Mistral and ReciTAL.</li> | |
| </ul> | |
| </div> | |
| </div> | |
| </div> | |
| <!--/ Text. --> | |
| <!-- Image --> | |
| <div class="has-text-centered" style="margin-bottom: 2rem;"> | |
| <img src="./static/images/A major national project.png" alt="Project logo" style="width: 300px;"> | |
| </div> | |
| <!--/ Image --> | |
| <!-- Text. --> | |
| <div class="columns is-centered"> | |
| <div class="column is-four-fifths"> | |
| <h2 class="title is-3">A four-step approach</h2> | |
| <div class="content has-text-justified"> | |
| <p> | |
| <strong>1. Developing Medical LLMs</strong> | |
| </p> | |
| <p> | |
| Building an open-source corpus of French medical text data to train, evaluate, and distribute multiple open-source medical language models. | |
| </p> | |
| <p> | |
| PARCOMED - PARTAGES Corpus of Open MEdical Documents available <a href="https://huggingface.co/datasets/HealthDataHub/PARCOMED" target="_blank">here</a> | |
| </p> | |
| <p> | |
| PARCOMED research only - PARTAGES Corpus of Open MEdical Documents (for research purpose only) available <a href="https://huggingface.co/datasets/HealthDataHub/PARCOMED" target="_blank">here</a> | |
| </p> | |
| <p> | |
| <strong>2. Creating an open database of fictional medical reports</strong> | |
| </p> | |
| <p> | |
| Creation and release as open data of a unique corpus of over 6,000 fictional medical reports, covering 20 specialties, including annotated reports for use cases. This effort involved more than 100 residents and young physicians and will be used in particular to train specialized models. | |
| </p> | |
| <p> | |
| PARHAF - an open French corpus of human-authored clinical reports of fictional patients available <a href="https://huggingface.co/datasets/HealthDataHub/PARHAF" target="_blank">here</a> | |
| </p> | |
| <p> | |
| <strong>3. Develop models for targeted use cases</strong> | |
| </p> | |
| <p> | |
| Using these resources, PARTAGES is developing seven specialized AI models designed for high-impact use cases in research, innovation, and the healthcare system. | |
| </p> | |
| <p> | |
| <strong>4. Establish a national federated evaluation platform</strong> | |
| </p> | |
| <p> | |
| Development of a sovereign federated evaluation platform, enabling the evaluation of algorithms on real-world data within a secure regulatory framework. It will be deployed in 20 healthcare facilities but may be used by any facility wishing to access it. | |
| </p> | |
| </div> | |
| </div> | |
| </div> | |
| <!--/ Text. --> | |
| <!-- Text. --> | |
| <div class="columns is-centered"> | |
| <div class="column is-four-fifths"> | |
| <h2 class="title is-3">Practical Use Cases for Healthcare</h2> | |
| <div class="content has-text-justified"> | |
| <p> | |
| PARTAGES addresses <strong>eight priority use cases</strong> focused on the analysis, structuring, and generation of medical reports: | |
| </p> | |
| <ul> | |
| <p><i class="fas fa-database" style="margin-right: 8px;"></i><strong> Data augmentation</strong> through the generation of fictitious reports</p> | |
| <p><i class="fas fa-comment-medical" style="margin-right: 8px;"></i><strong> Automatic pseudonymization</strong> of medical reports</p> | |
| <p><i class="fas fa-laptop-code" style="margin-right: 8px;"></i><strong> Automated medical coding (DIM</strong>) based on medical reports</p> | |
| <p><i class="fas fa-brain" style="margin-right: 8px;"></i><strong> Automatic summarization</strong> of medical reports</p> | |
| <p><i class="fas fa-hand-holding-medical" style="margin-right: 8px;"></i><strong> Generation of clinical</strong> cases for medical training</p> | |
| <p><i class="fas fa-dna" style="margin-right: 8px;"></i><strong> Identification of tumor biomarkers</strong> in oncology</p> | |
| <p><i class="fas fa-solid fa-pills" style="margin-right: 8px;"></i><strong> Analysis of treatment response</strong> in oncology</p> | |
| <p><i class="fas fa-bacteria" style="margin-right: 8px;"></i><strong> Automatic detection in infectious diseases</strong>, particularly to combat antibiotic resistance</p> | |
| </ul> | |
| </div> | |
| </div> | |
| </div> | |
| <!--/ Text. --> | |
| <div class="content"> | |
| <h2 class="title is-3">An organization structured into 8 work packages</h2> | |
| <ul> | |
| <li> | |
| <i class="fas fa-database" style="margin-right: 8px;"></i> | |
| <strong>WP1 – Data collection and governance:</strong> Structuring, accessing and securing health data. | |
| </li> | |
| <p> | |
| The first work package is dedicated to the overall coordination of the PARTAGES project, as well as the dissemination and promotion of its results. It ensures sound project governance, consistency among the various work packages, and adherence to the timeline, scientific objectives, and regulatory requirements. | |
| </p> | |
| <li> | |
| <i class="fas fa-cogs" style="margin-right: 8px;"></i> | |
| <strong>WP2 – Data preprocessing and harmonization:</strong> Standardizing and preparing datasets for model training. | |
| </li> | |
| <p> | |
| The objective of this second phase is to establish a rigorous methodology for the production and use of project data (primarily fictitious medical records) and to ensure the quality control of all raw datasets used to train the project’s models. It also supports data quality control at the healthcare facility level for model evaluation. | |
| </p> | |
| <li> | |
| <i class="fas fa-brain" style="margin-right: 8px;"></i> | |
| <strong>WP3 – Foundation model development:</strong> Designing and training large-scale models for health data. | |
| </li> | |
| <p> | |
| Work Package 3 is responsible for developing and implementing a common evaluation methodology for foundation models and all use cases, as well as for analyzing the evaluation results. | |
| </p> | |
| <li> | |
| <i class="fas fa-vial" style="margin-right: 8px;"></i> | |
| <strong>WP4 – Use cases in oncology:</strong> Identification of tumor biomarkers and clinical applications. | |
| </li> | |
| <p> | |
| The objective of the fourth work package is to develop all the foundational models that will be used by the use cases, including the fine-tuning of the general-purpose generative LLM for the French-language medical domain, as well as the development of BERT-style encoder models (Bidirectional Encoder Representations from Transformers). | |
| </p> | |
| <li> | |
| <i class="fas fa-bacteria" style="margin-right: 8px;"></i> | |
| <strong>WP5 – Use cases in infectious diseases:</strong> Automatic detection and support for combating antibiotic resistance. | |
| </li> | |
| <p> | |
| Lot No. 5 focuses on establishing the necessary technical infrastructure, including the creation, adaptation, and documentation of the federated validation platform, in which each partner healthcare facility serves as a node. | |
| </p> | |
| <li> | |
| <i class="fas fa-chart-line" style="margin-right: 8px;"></i> | |
| <strong>WP6 – Evaluation and benchmarking:</strong> Assessing model performance, robustness and reproducibility. | |
| </li> | |
| <p> | |
| The goal of the sixth batch is to manage the recruitment and supervision of healthcare professionals (senior residents and junior physicians) for the creation of a corpus of 5,000 fictional patient records and for annotation tasks. | |
| </p> | |
| <li> | |
| <i class="fas fa-network-wired" style="margin-right: 8px;"></i> | |
| <strong>WP7 – Deployment and infrastructure:</strong> Integration into secure environments and operationalization. | |
| </li> | |
| <p> | |
| Lot No. 7 oversees the project's legal matters, including the implementation of the contractual framework and the monitoring of work related to the local use of healthcare facility reports. | |
| </p> | |
| <li> | |
| <i class="fas fa-users" style="margin-right: 8px;"></i> | |
| <strong>WP8 – Dissemination and collaboration:</strong> Promoting open science, sharing resources and engaging stakeholders. | |
| </li> | |
| <p> | |
| This final phase covers the development of models for the specific use cases identified above. | |
| </p> | |
| </ul> | |
| </div> | |
| <div class="content"> | |
| <h2 class="title is-3">PARTAGES stakeholders</h2> | |
| <h3 class="title is-5"> | |
| <i class="fas fa-university" style="margin-right: 8px;"></i> | |
| Academic and research institutions | |
| </h3> | |
| <ul> | |
| <li>LISN</li> | |
| <li>GREYC</li> | |
| <li>LIG</li> | |
| <li>LIMICS</li> | |
| <li>LIA</li> | |
| <li>LS2N</li> | |
| <li>LIS</li> | |
| <li>UNESS</li> | |
| <li>Bordeaux Population Health</li> | |
| <li>CNRS</li> | |
| <li>INRIA</li> | |
| </ul> | |
| <h3 class="title is-5"> | |
| <i class="fas fa-hospital" style="margin-right: 8px;"></i> | |
| Hospitals and healthcare institutions | |
| </h3> | |
| <ul> | |
| <li>AP-HP</li> | |
| <li>Centre Léon Bérard</li> | |
| <li>CHU de Rouen</li> | |
| <li>GCS HOURAA</li> | |
| <li>Institut Gustave Roussy</li> | |
| <li>Institut Curie</li> | |
| <li>Hôpitaux Saint-Joseph & Marie-Lannelongue</li> | |
| <li>Hôpital Foch</li> | |
| <li>CHU de Toulouse</li> | |
| <li>CHU de Bordeaux</li> | |
| <li>CHU Amiens-Picardie</li> | |
| <li>CHU Brest</li> | |
| <li>CHU de Lille</li> | |
| <li>CHRU de Nancy</li> | |
| <li>CHU de La Réunion</li> | |
| </ul> | |
| <h3 class="title is-5"> | |
| <i class="fas fa-building" style="margin-right: 8px;"></i> | |
| Industry and private partners | |
| </h3> | |
| <ul> | |
| <li>ELSAN</li> | |
| <li>Ramsay Santé</li> | |
| <li>reciTAL</li> | |
| <li>Mistral AI</li> | |
| </ul> | |
| <h3 class="title is-5"> | |
| <i class="fas fa-landmark" style="margin-right: 8px;"></i> | |
| Public coordination | |
| </h3> | |
| <ul> | |
| <li>Health Data Hub</li> | |
| </ul> | |
| </div> | |
| <footer class="footer"> | |
| <div class="container"> | |
| <div class="content has-text-centered"> | |
| <a class="icon-link" target="_blank" | |
| href="./static/videos/nerfies_paper.pdf"> | |
| <i class="fas fa-file-pdf"></i> | |
| </a> | |
| <a class="icon-link" href="https://github.com/keunhong" target="_blank" class="external-link" disabled> | |
| <i class="fab fa-github"></i> | |
| </a> | |
| </div> | |
| <div class="columns is-centered"> | |
| <div class="column is-8"> | |
| <div class="content"> | |
| <p> | |
| This website is licensed under a <a rel="license" target="_blank" | |
| href="http://creativecommons.org/licenses/by-sa/4.0/">Creative | |
| Commons Attribution-ShareAlike 4.0 International License</a>. | |
| </p> | |
| <p> | |
| This means you are free to borrow the <a target="_blank" | |
| href="https://github.com/nerfies/nerfies.github.io">source code</a> of this website, | |
| we just ask that you link back to this page in the footer. | |
| Please remember to remove the analytics code included in the header of the website which | |
| you do not want on your website. | |
| </p> | |
| </div> | |
| </div> | |
| </div> | |
| </div> | |
| </footer> | |
| </body> | |
| </html> | |