BEIJING, Aug. 4, 2022 /PRNewswire/ — The 5G+AI Work Group of IMT-2020 (5G) Promotion Group has officially launched The third Wireless Communication AI Competition on July 25, 2022(Beijing time) on DataFountain (a leading Big Data & AI competition platform).

The competition is open to all the people around the world, regardless of age or nationality. All colleges and universities, scientific research institutions, enterprises, maker teams, individuals, etc. can log on to the official website to register for the competition.

COMPETITION BACKGROUND

5G+AI has become the most popular research area for the development of wireless communications. Competition becomes an unprecedented way to promote the integration of 5G and AI by involving academic and industrial R&D to solve typical wireless problems with AI/ML tools.

The 5G+AI Work Group of IMT-2020 (5G) Promotion Group is committed to conducting research on the requirements, theories and technologies of deep integration of 5G and AI, and promoting the international standardization and industrialization process of 5G and AI integration. Wireless Communication AI Competition (WAIC) is hosted by The 5G+AI Work Group of IMT-2020 (5G) Promotion Group for researchers to contribute to the integration of wireless communications and AI, with the theme of “AI Enlightens Wireless Communication”. A continue influence is expected for the WAIC to attract academic and industry contribution to the integration of 5G and AI.

COMPETITION TRACK IN THE 3rd WAIC

Track 1: AI-based Joint Channel Estimation and Channel State Information Feedback

Track 2: AI-based High-Precision Positioning

COMPETITION TASK

Track 1: AI-based Joint Channel Estimation and Channel State Information Feedback

Organizer: CAICT  OPPO

Task: In this competition, we consider to make use of the information compression, nonlinear recovery and end-to-end optimization capability of AI, to take the received pilot signal (CSI-RS) on small number of time-frequency resources at user equipment (UE) side as the input, and strive to realize the recovery of channel state information as accurately as possible at base station (BS) side, so that enabling the construction of high-quality wireless communication link.

Specifically, the characteristics of this competition are as follows:

a. Following the practical requirements of wireless communication system, the competition focuses on the improvement of CSI feedback performance based on received pilot signal with limited number.

b. The solution design with two different pilot densities (sufficient and insufficient pilot) are required, and the performance will be evaluated separately.

c. In the traditional communication module partitioning, channel estimation and CSI feedback are independently designed and optimized. With the empowerment of AI, there are actually a variety of potential implementation methods to improve the overall performance, such as based on integrated design or modular design. Participants can freely select the most appropriate implementation method to achieve the best performance.

Track 2: AI-based High-Precision Positioning

Organizer: CAICT, Huawei and vivo

Task: This competition includes three cases, which examine the contestants’ capabilities of using AI to obtain high-precision positioning accuracy. In general, several AI models need to be designed and trained by considering the channel knowledge, i.e., channel impulse response, as the network input. The output is indeed the user coordinate. Contestants need to optimize the network architecture as well as the training procedure, so as to get a more accurate inferencing result which will be compared with the ground-truth label to indicate the score.

Specifically, the characteristics of this competition are listed as follows:

a. The provided AI model(s) needs to have good scalability and generalization capabilities to handle different testing scenarios, i.e., spatial consistency differences

b. The provided AI model(s) needs to work well when the distribution of the input data is shifting, i.e., the number of available BSs might differ

c. The provided AI model(s) needs to have the capabilities to work well when the percentage of labeled data samples is relatively small. In this case, unlabeled data needs to be well utilized to improve the positioning performance.

COMPETITION SCHEDULE

The competition will be held online, and technology sharing and awarding will be carried out offline. Contestants should register, team up, and submit the model for evaluation on the official platform DataFountain. After the online competition, the top ten teams or individuals whose submitted models have been reviewed need to participate in the seminar and attend the award ceremony.

The schedule is as follows:

–      July 25, 2022: The competition tasks are released, contestants can log in to the official website for registering:https://sourl.cn/qfsGMJ

–      August 1, 2022: The dataset is released. Contestants can log in to the official website to download the dataset. At the same time, the competition officially enters the evaluation stage. Players can submit the model files online to the competition platform. Each team can submit up to 5 times a day. The evaluation system will automatically evaluate the score.

–      September 25, 2022 12:00 (Beijing time): Deadline for registration and team formation.

–      September 28, 2022 12:00 (Beijing time): The evaluation platform is closed, and the score list is locked.

–      September 29-October 31, 2022: After reviewing the models of the top ten teams, the teams will be awarded the Golden, Silver, Bronze and the Winning prize in descending order of the evaluation results. The winning teams must participate in the seminar and the award ceremony. Teams that do not participate will abandon the competition by default.

–      November 2022 [to be confirmed]: Seminar and award ceremony.

Bonus (for Both Track)

Golden Prize: 300,000 RMB for 1 team

Silver Prize: 50,000 RMB (per team) for 2 teams

Bronze Prize: 20,000 RMB (per team) for 3 teams

Winning Prize: 10,000 RMB (per team) for 4 teams

Official Competition websitehttps://sourl.cn/qfsGMJ

Click here to learn more and sign up.