Pozarren aurkezten dugu Latxa eredu irekien familia, euskarazko hizkuntza eredurik handiena eta hoberena duena.
Read more
EDHIA project is framed within a coordinated effort to advance the early detection of various high-impact diseases through the application of natural language processing (NLP) and artificial intelligence (AI) on medical documents. The ultimate goal is to develop tools that can process large volumes of clinical data—such as Electronic Health Records (EHRs), clinical notes, and scientific literature—to identify early risk factors and support healthcare professionals in early diagnosis and intervention.
Objective and Scope:
Mental Health: Improving early detection of mental health issues, especially those that are often underreported or stigmatized, such as suicidal tendencies. The project leverages NLP techniques to process patient records and flag potential risks that might otherwise go unnoticed by physicians.
HIV Detection: Enhancing the diagnosis of HIV infections through NLP by identifying missed opportunities for testing and intervention in the patient records, supporting the World Health Organization’s 95-95-95 goal for HIV management by 2030.
Rare Diseases: Focusing on improving the quality of life for patients with rare diseases (RDs), particularly children. This involves identifying the connection between congenital malformations and their mental health evolution, along with other social determinants of health.
Cardiovascular Complications: Predicting risk factors related to cardiovascular diseases, especially after an initial episode of Atrial Fibrillation, using AI and NLP on structured data such as electrocardiogram reports and unstructured clinical notes.
Duration and Collaboration:
The project has a duration of 36 months and involves collaboration between several leading academic institutions:
-
- HiTZ (Basque Center for Language Technology): Focuses on language models and corpora annotation.
-
- UNED (National University of Distance Education): Leads the effort on developing computational tools for NLP in the medical domain.
-
Metodology:
The project will use a combination of structured and unstructured data, applying advanced NLP techniques such as temporal pattern detection, medical ontology enrichment, and language model fine-tuning to support the early detection tasks. These tools will be adapted to multilingual environments, making the solutions applicable across various linguistic contexts.
This collaboration ensures the generalization of the developed systems across different healthcare datasets, aiming to provide scalable solutions that can be applied across hospitals and healthcare systems.
Financial Support: Department of Culture and Language Policy
