Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions
Deep learning-based algorithms have led to tremendous progress over the last years, but they face a bottleneck as their optimal development highly relies on access to large datasets. To mitigate this limitation, cross-silo federated learning has emerged as a way to train collaborative models among m...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-01-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/13/2/919 |
| _version_ | 1797446604262735872 |
|---|---|
| author | Laëtitia Launet Yuandou Wang Adrián Colomer Jorge Igual Cristian Pulgarín-Ospina Spiros Koulouzis Riccardo Bianchi Andrés Mosquera-Zamudio Carlos Monteagudo Valery Naranjo Zhiming Zhao |
| author_facet | Laëtitia Launet Yuandou Wang Adrián Colomer Jorge Igual Cristian Pulgarín-Ospina Spiros Koulouzis Riccardo Bianchi Andrés Mosquera-Zamudio Carlos Monteagudo Valery Naranjo Zhiming Zhao |
| author_sort | Laëtitia Launet |
| collection | DOAJ |
| description | Deep learning-based algorithms have led to tremendous progress over the last years, but they face a bottleneck as their optimal development highly relies on access to large datasets. To mitigate this limitation, cross-silo federated learning has emerged as a way to train collaborative models among multiple institutions without having to share the raw data used for model training. However, although artificial intelligence experts have the expertise to develop state-of-the-art models and actively share their code through notebook environments, implementing a federated learning system in real-world applications entails significant engineering and deployment efforts. To reduce the complexity of federation setups and bridge the gap between federated learning and notebook users, this paper introduces a solution that leverages the Jupyter environment as part of the federated learning pipeline and simplifies its automation, the Notebook Federator. The feasibility of this approach is then demonstrated with a collaborative model solving a digital pathology image analysis task in which the federated model reaches an accuracy of 0.8633 on the test set, as compared to the centralized configurations for each institution obtaining 0.7881, 0.6514, and 0.8096, respectively. As a fast and reproducible tool, the proposed solution enables the deployment of a cross-country federated environment in only a few minutes. |
| first_indexed | 2024-03-09T13:42:59Z |
| format | Article |
| id | doaj.art-1efbf6ae12c34873b162e6bc3f1d2c77 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-09T13:42:59Z |
| publishDate | 2023-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-1efbf6ae12c34873b162e6bc3f1d2c772023-11-30T21:03:30ZengMDPI AGApplied Sciences2076-34172023-01-0113291910.3390/app13020919Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed InstitutionsLaëtitia Launet0Yuandou Wang1Adrián Colomer2Jorge Igual3Cristian Pulgarín-Ospina4Spiros Koulouzis5Riccardo Bianchi6Andrés Mosquera-Zamudio7Carlos Monteagudo8Valery Naranjo9Zhiming Zhao10CVBLab, Instituto Universitario de Investigación en Tecnología Centrada en el Ser Humano (HUMAN-Tech), Universitat Politècnica de València, 46022 Valencia, SpainMultiscale Networked Systems, University of Amsterdam, 1098 XH Amsterdam, The NetherlandsCVBLab, Instituto Universitario de Investigación en Tecnología Centrada en el Ser Humano (HUMAN-Tech), Universitat Politècnica de València, 46022 Valencia, SpainInstituto de Telecomunicaciones y Aplicaciones Multimedia (ITEAM), Departamento de Comunicaciones, Universitat Politècnica de València, 46022 Valencia, SpainCVBLab, Instituto Universitario de Investigación en Tecnología Centrada en el Ser Humano (HUMAN-Tech), Universitat Politècnica de València, 46022 Valencia, SpainMultiscale Networked Systems, University of Amsterdam, 1098 XH Amsterdam, The NetherlandsMultiscale Networked Systems, University of Amsterdam, 1098 XH Amsterdam, The NetherlandsPathology Department, Hospital Clínico Universitario de Valencia, Universidad de Valencia, 46010 Valencia, SpainPathology Department, Hospital Clínico Universitario de Valencia, Universidad de Valencia, 46010 Valencia, SpainCVBLab, Instituto Universitario de Investigación en Tecnología Centrada en el Ser Humano (HUMAN-Tech), Universitat Politècnica de València, 46022 Valencia, SpainMultiscale Networked Systems, University of Amsterdam, 1098 XH Amsterdam, The NetherlandsDeep learning-based algorithms have led to tremendous progress over the last years, but they face a bottleneck as their optimal development highly relies on access to large datasets. To mitigate this limitation, cross-silo federated learning has emerged as a way to train collaborative models among multiple institutions without having to share the raw data used for model training. However, although artificial intelligence experts have the expertise to develop state-of-the-art models and actively share their code through notebook environments, implementing a federated learning system in real-world applications entails significant engineering and deployment efforts. To reduce the complexity of federation setups and bridge the gap between federated learning and notebook users, this paper introduces a solution that leverages the Jupyter environment as part of the federated learning pipeline and simplifies its automation, the Notebook Federator. The feasibility of this approach is then demonstrated with a collaborative model solving a digital pathology image analysis task in which the federated model reaches an accuracy of 0.8633 on the test set, as compared to the centralized configurations for each institution obtaining 0.7881, 0.6514, and 0.8096, respectively. As a fast and reproducible tool, the proposed solution enables the deployment of a cross-country federated environment in only a few minutes.https://www.mdpi.com/2076-3417/13/2/919federated learningJupyter notebookmedical image analysiscollaborative modelscloud environmentdistributed medical applications |
| spellingShingle | Laëtitia Launet Yuandou Wang Adrián Colomer Jorge Igual Cristian Pulgarín-Ospina Spiros Koulouzis Riccardo Bianchi Andrés Mosquera-Zamudio Carlos Monteagudo Valery Naranjo Zhiming Zhao Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions Applied Sciences federated learning Jupyter notebook medical image analysis collaborative models cloud environment distributed medical applications |
| title | Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions |
| title_full | Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions |
| title_fullStr | Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions |
| title_full_unstemmed | Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions |
| title_short | Federating Medical Deep Learning Models from Private Jupyter Notebooks to Distributed Institutions |
| title_sort | federating medical deep learning models from private jupyter notebooks to distributed institutions |
| topic | federated learning Jupyter notebook medical image analysis collaborative models cloud environment distributed medical applications |
| url | https://www.mdpi.com/2076-3417/13/2/919 |
| work_keys_str_mv | AT laetitialaunet federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT yuandouwang federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT adriancolomer federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT jorgeigual federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT cristianpulgarinospina federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT spiroskoulouzis federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT riccardobianchi federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT andresmosquerazamudio federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT carlosmonteagudo federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT valerynaranjo federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions AT zhimingzhao federatingmedicaldeeplearningmodelsfromprivatejupyternotebookstodistributedinstitutions |