The Open-Source Revolution of AliBERT in French Biomedical AI

2024, 21 February

| 2 min read|
Grégoire Dugast
Thumbnail of the article about AliBERT available in open source

Open-Source Contribution

Developed by Quinten in 2022, AliBERT is a specialized model focusing on the French biomedical language. One version has been released as an open-source tool on the Huggingface platform, marking a significant contribution to the field of natural language processing (NLP) in healthcare. This approach of providing open access fosters a culture of innovation and collaboration across the global research community, to better serve the needs of healthcare professionals and researchers.

AliBERT’s Core Applications

AliBERT demonstrates state-of-the-art performance in several key areas, including but not limited to:

  • Extracting biomedical concepts: to support research and clinical practice, especially in oncology.
  • Detecting medication dosages: to ensure accurate patient care.
  • Pseudonymizing patient reports: to maintain confidentiality.
  • Biomedical Term Codification (ICD-10): To streamline medical records for easier access and analysis.

The acknowledgment of AliBERT’s contributions to the biomedical field was further solidified by a publication at the prestigious ACL (Association for Computational Linguistics) conference in 2023, highlighting its innovative applications and impact.

Ongoing Development and Enhancement

Meanwhile, Quinten’s DataLab is committed to the ongoing development and diversification of the AliBERT model ecosystem. This dedication to continuous improvement not only aims to augment AliBERT’s functionalities and discover new healthcare applications but also ensures access to the most sophisticated versions of AliBERT, specifically refined for advanced uses. By persistently enhancing AliBERT, we maintain its status as a leading-edge NLP technology solution, adeptly addressing the complex demands of French biomedical language processing.


latest articles