CAIDAS Contributions on Graph Neural Networks Accepted at NeurIPS 2024
10/29/2024Two research projects from the CAIDAS-Chair of Machine Learning for Complex Networks will be featured at NeurIPS 2024. Led by Prof. Dr. Ingo Scholtes and his team, the studies explore new applications of graph neural networks for predicting temporal centralities and improving community detection. These findings will be presented in December in Vancouver, Canada.
We are happy to announce that two works from the CAIDAS-Chair of Machine Learning for Complex Networks were accepted for the research track of the Conference on Neural Information Processing Systems (NeurIPS) 2024, which is one of the world’s primary conferences on deep learning.
A first work by PhD student Franziska Heeg and Prof. Dr. Ingo Scholtes uses time-aware graph neural networks to predict temporal centralities. In a second work by postdoc Dr. Christopher Blöcker, PhD student Chester Tan and Prof. Dr. Ingo Scholtes shows how graph neural networks can be used to minimize the information-theoretic MapEquation to address community detection in graphs.
Both works will be presented at NeurIPS in December 2024 in Vancouver, Canada.
