Authors: José Alberto Benítez-Andrades, María Teresa García-Ordás, Mayra Russo, Ahmad Sakor, Luis Daniel Fernandes Rotger, Maria-Esther Vidal

Published on: February 08, 2024

Impact Score: 8.38

Arxiv code: Arxiv:2402.05536

Summary

• What is new: Introduces a hybrid approach combining community-maintained knowledge graphs with deep learning for understanding social media posts in the health domain.
• Why this is important: The challenge of Artificial Intelligence in understanding the context of brief text posts on social networks, especially regarding health-related issues.
• What the research proposes: A novel method that uses advanced entity recognizers and linkers to connect short post entities to knowledge graphs and employs knowledge graph embeddings (KGEs) and contextualized word embeddings to create context-based representations of posts.
• Results: Improved predictive models’ reliability in identifying posts related to eating disorders, aiding early diagnosis for healthcare providers.

Technical Details

Technological frameworks used: Falcon 2.0 for entity recognition and linking; deep learning frameworks for modeling

Models used: BERT for contextualized word embeddings; Knowledge Graph Embeddings for integrating structured knowledge

Data used: 2,000 tweets about eating disorders

Potential Impact

Healthcare providers, mental health specialists, social media platforms, and Artificial Intelligence developers in the medical and mental health domain.

Want to implement this idea in a business?

We have generated a startup concept here: HealthGraphAI.