Expert recommendations based on link prediction during the COVID-19 outbreak.

Area of Science:

• Computer Science
• Artificial Intelligence
• Bioinformatics

Background:

• The COVID-19 pandemic has led to a surge in infections and deaths globally.
• Despite research efforts, effective vaccines and therapeutics are still under development, with limited interdisciplinary cooperation.
• Information silos and redundant research hinder progress in finding COVID-19 solutions.

Purpose of the Study:

• To analyze features of expert cooperation networks to identify potential collaborators.
• To develop a novel expert recommendation model to encourage interdisciplinary and multinational research cooperation.
• To accelerate the development of therapeutic drugs and vaccines for COVID-19.

Main Methods:

• Comprehensive analysis of expert cooperation network features, including graph structure, context attributes, and sequential co-occurrence probability.
• Construction of a novel Graph Neural Network (GNN) + Long Short-Term Memory (LSTM) + Generative Adversarial Network (GAN) model.
• Application of link prediction techniques within the developed GNN+LSTM+GAN model to recommend expert collaborators.

Main Results:

• The study provides a framework for analyzing expert networks and identifying key features for collaboration.
• The proposed GNN+LSTM+GAN model demonstrates potential for predicting and recommending expert connections.
• The model aims to facilitate the formation of cooperative relationships among researchers in social networks.

Conclusions:

• Enhanced expert collaboration is crucial for accelerating the development of COVID-19 vaccines and therapeutics.
• The developed AI model offers a promising approach to overcoming research fragmentation and promoting cross-field cooperation.
• Facilitating multinational and cross-field expert partnerships is significant for global health challenges.