AI/ML in 5G Challenge

Applying machine learning in communications networks.

The ITU AI/ML in 5G Challenge aims to provide a platform for collaboratively addressing the problems in applying AI/ML in future networks including 5G. The 2021 edition of the Challenge connected more than 1600 participants from 82 countries, with industry and academia solving real-world problems using AI/ML in networks. 

The Challenge offers carefully curated problem statements, a mix of real-world and simulated data, technical webinars, mentoring, and hands-on- sessions. Teams participating in the Challenge enable, create, train and deploy ML models for communication networks. This enables participants to not only showcase their talent, test their concepts on real data and real-world problems, and compete for global recognition, but also enter the world of ITU standards by mapping their solutions to our specifications.  

The ITU AI/ML in 5G Challenge aims to provide a platform for collaboratively addressing the problems in applying AI/ML in future networks including 5G. The 2021 edition of the Challenge connected more than 1600 participants from 82 countries, with industry and academia solving real-world problems using AI/ML in networks. 

The Challenge offers carefully curated problem statements, a mix of real-world and simulated data, technical webinars, mentoring, and hands-on- sessions. Teams participating in the Challenge enable, create, train and deploy ML models for communication networks. This enables participants to not only showcase their talent, test their concepts on real data and real-world problems, and compete for global recognition, but also enter the world of ITU standards by mapping their solutions to our specifications.  

Compute platform

ITU provides a state-of-the-art, free-of-charge compute platform to participants of the Challenge who do not have adequate access to compute in their respective institutions. The compute platform will provide participants with access to:

  • Free GPUs and CPUs
  • Hosted Jupyter notebook server
  • Python kernel
  • Pre-installed machine learning packages, e.g. PyTorch and Tensorflow

In some of the problem statements, a baseline or reference solution may be offered which may include implementations using Jupyter notebooks.

ML/5G Challenge problem statements

Federated Traffic Prediction

Usage of federated learning tools to predict traffic in cellular networks from real measurements

Curated by CTTC

Graph Neural Networking Challenge

Exploring a data-centric approach to develop Network Digital Twins by producing a training dataset that results in better performance of the target GNN model

Curated by BNN-UPC

Beam Classification with DeepBeam

To create a classifier for transmit beams out of a codebook using neural networks with the DeepBeam Dataset

Curated by Northeastern University

Depth Map Estimation in 6G

Propose an AI/ML algorithm that reconstructs the depth map of an indoor environment from mmWave MIMO channel impulse responses and depth map representation

Curated by NIST (National Institute of Standards and Technology)

Network failure prediction

Create AI/ML models for predicting network failures using time-series data consisting of thousands of metrics in 5G core network

Curated by KDDI

Location Estimation Using RSSI

To develop an AI/ML-based localization algorithm/technique that accurately estimate the position of a receiver based on RSS information

Curated by RISING

Behavioral modeling for energy efficiency

To bring AI algorithms into behavioral modeling of power amplifiers, helpful for developing highly efficient 5G wireless communication systems

Curated by ZTE

Throughput Prediction in Wi-Fi networks

Use of machine learning to predict achievable throughput that a subset of APs transmitting at the same time can achieve

Curated by UPF

Data Generation using GANs

Generating synthetic observability data using GANs for Telco-Cloud Infrastructure Metrics to solve the problem of dataset unavailability for AI/ML researchers

Curated by LF Networking

Build your own Closed loop

Collaboratively create a crowdsourced, baseline representation for AN closed loops (controllers) as a proof of concept

Curated by FG-AN

Beam Prediction Challenge

To design machine learning-based models that can adapt to and perform accurate sensing-aided beam prediction at an entirely new location

Curated by Arizona State University

Classification of network users

To classify each user and accurately distinguish between users with bad experience and users with good experience using ML/DL methods

Curated by ZTE

ITU AI/ML 5G Challenge Timeline

February – May 2022

June – October 2022

December 2022

Curation Phase

Competition Phase

Grand Challenge Finale (online) 

Registration: ~ 21 October 2022

Submission of solutions

  • 28 October

Evaluation of solutions: October 2022

Final Ranking: 31 October 2022

Preparation of reports: October – November 2022

November 2022 – Judges Panel evaluates the best solutions from Competition Phase

29 Nov – 01 Dec 2022 – Best solutions pitch in a 3-day event end of to determine the finalists

14 December 2022 – Grand Challenge Finale

ITU AI/ML 5G Challenge Timeline

February – May 2022

Curation Phase

June – October 2022

Competition Phase

  • 21 October 2022 – Registration
  • 28 October – Submission of solutions
  • October 2022 – Evaluation of solutions
  • October – November 2022 – Preparation of reports

December 2022

Grand Challenge Finale (online) 

  • November 2022 – Judges Panel evaluates the best solutions from Competition Phase
  • Best solutions pitch in a 3-day event 29 Nov – 01 Dec 2022 to determine the finalists
  • 14 December 2022 – Grand Challenge Finale
Related sessions
15 June 2022
15:00 - 16:30 CEST Geneva | 21:00-22:30 CST, Beijing | 09:00-10:30 EST, New York
Jelena Senic (National Institute of Standards and Technology (NIST)), Raied Caromi (National Institute of Standards and Technology, CTL, Wireless and Network Division), Steve Blandino (National Institute of Standards and Technology (NIST))
1 July 2022
15:00 - 16:30 CEST, Geneva | 09:00-10:30 EDT, New York | 21:00-22:30 CST, Beijing
Thomas Basikolo‬ (International Telecommunication Union (ITU)), Vishnu Ram OV (International Telecommunication Union (ITU)), Kai Lu (China Mobile Research Institute)...
6 July 2022
09:00 - 11:00 CEST, Geneva | 16:00-18:00 JST, Tokyo | 03:00-05:00 EST, New York
Akihiro Nakao (University of Tokyo), Anan Sawabe (NEC Corporation), Koichi Adachi (University of ElectroCommunications)...
15 July 2022
14:00 - 15:10 CEST, Geneva | 08:00-09:10 EDT, New York | 20:00-21:10 CST, Beijing
Thomas Basikolo‬ (International Telecommunication Union (ITU)), Vishnu Ram OV (International Telecommunication Union (ITU)), Yansha Deng (King's College London)
AI for Good ML/5G partners

Sponsor

Hosts of problem statements

Technical partner

Artificial Intelligence Industry Alliance
Artificial Intelligence Industry Alliance

Sponsorship Inquiries​


    Testimonials

    Your Gateway to #AIforGood in your inbox!

    Join our 60K+ global community to receive insights & invites to AI for Good! Just one email each week, and zero spam.