Skip to main content Skip to navigation

The Graph Neural Networking Challenge: A Worldwide Competition for Education in AI/ML for Networks

Project Overview

The document outlines the Graph Neural Networking Challenge 2020, which successfully engaged over 1300 participants worldwide in leveraging machine learning (ML) for networking challenges, particularly through the application of Graph Neural Networks (GNN). The initiative aimed to not only educate students and professionals but also to foster collaboration and innovation in predicting network performance metrics using ML techniques. By providing a competitive platform, the challenge facilitated the development of practical solutions, generated valuable educational resources, and underscored significant challenges in the realm of network modeling. Overall, the challenge demonstrated the potential of generative AI and ML in enhancing educational experiences and addressing complex problems in the networking field.

Key Applications

Graph Neural Networking Challenge 2020

Context: Educational competition for students, researchers, and professionals in AI/ML for networks

Implementation: Participants were provided with datasets, a baseline model (RouteNet), and resources to develop solutions over a six-month period.

Outcomes: Created educational resources (datasets, API, tutorials), engaged over 1300 participants, and established benchmarks for network modeling.

Challenges: Difficulty in obtaining publicly available datasets and ensuring solutions comply with competition rules.

Implementation Barriers

Data Availability

Lack of public datasets for ML applications in networking, making it hard to train and evaluate models.

Proposed Solutions: ML competitions can help establish relevant open datasets for benchmarking existing solutions.

Compliance Verification

Ensuring that all submitted solutions comply with the competition rules, especially regarding the use of neural networks.

Proposed Solutions: Manual review of top solutions and potential use of sandboxes for participants to run their solutions.

Project Team

José Suárez-Varela

Researcher

Miquel Ferriol-Galmés

Researcher

Albert López

Researcher

Paul Almasan

Researcher

Guillermo Bernárdez

Researcher

David Pujol-Perich

Researcher

Krzysztof Rusek

Researcher

Loïck Bonniot

Researcher

Christoph Neumann

Researcher

François Schnitzler

Researcher

François Taïani

Researcher

Martin Happ

Researcher

Christian Maier

Researcher

Jia Lei Du

Researcher

Matthias Herlich

Researcher

Peter Dorfinger

Researcher

Nick Vincent Hainke

Researcher

Stefan Venz

Researcher

Johannes Wegener

Researcher

Henrike Wissing

Researcher

Bo Wu

Researcher

Shihan Xiao

Researcher

Pere Barlet-Ros

Researcher

Albert Cabellos-Aparicio

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: José Suárez-Varela, Miquel Ferriol-Galmés, Albert López, Paul Almasan, Guillermo Bernárdez, David Pujol-Perich, Krzysztof Rusek, Loïck Bonniot, Christoph Neumann, François Schnitzler, François Taïani, Martin Happ, Christian Maier, Jia Lei Du, Matthias Herlich, Peter Dorfinger, Nick Vincent Hainke, Stefan Venz, Johannes Wegener, Henrike Wissing, Bo Wu, Shihan Xiao, Pere Barlet-Ros, Albert Cabellos-Aparicio

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies