On May 17, the first 6G intelligent wireless communication system competition sharing meeting and award ceremony was successfully held in Shanghai, announcing the successful conclusion of the competition.
Group photo at the awards ceremony
This competition was put forward in a timely manner during the critical development stage of 6G research. It aims to promote 6G vision, advanced technologies and concepts to all sectors of society, attract outstanding talents from the whole society, systematically and multi-anglely analyze and study the impact of AI on future wireless communication systems and solutions to key problems, with a view to comprehensively promoting the technological breakthrough of smart 6G and building the industry's first-mover advantage.
During the three-month competition schedule, thousands of teams and players signed up for the competition. In the end, a total of 32 teams ranked among the top of the leaderboard for 3 competition questions and won competition awards.
The first 6G intelligent wireless communication system competition sharing meeting and award ceremony
At the sharing meeting, special guests from China Academy of Information and Communications Technology, Huawei, OPPO, Zhejiang University, Shanghai Science and Technology Commission and other organizations delivered speeches. Experts and scholars in the field of wireless communication highly recognized the value and influence of the event, affirmed the works and technical capabilities of the contestants, and expressed their expectation for the event to continuously explore new fields and new directions, and make more positive contributions to the development of the 6G technology industry.
Special guest speech
Contest Highlights
Accelerate the high-quality development of 6G in line with the trend
Responding to the national strategy of "innovation-driven development" and focusing on the needs of technological innovation, the competition focuses on 6G research, creates an atmosphere of innovation and creativity in academic and industrial R&D through the competition, promotes the process of 6G industrialization, and promotes the integrated development of innovative elements such as industry, academics, and talents.
Authoritative organization, opening new competitions in the 6G field
The competition organizer IMT-2030 (6G) promotion group, the organizer China Academy of Information and Communications Technology, Huawei, OPPO, Zhejiang University, Shanghai Science and Technology Commission and other organizations have always been committed to opening 6G new communication competitions, actively exploring professional talents, and comprehensively promoting the technological breakthrough and industrial implementation of smart 6G.
High bonuses, spending millions to recruit talents
There are three competition topics set up in the competition, with a total prize pool of 1.5 million yuan. High incentives are used to stimulate 6G enthusiasts to participate. It is hoped that outstanding talents and advanced technologies in the fields of artificial intelligence and wireless communications will be brought together, so as to accelerate the pace of my country's self-reliance and self-improvement in science and technology and the construction of a technologically powerful country. Add benefits.
Simple competition system, the first evaluation will be the final
The objective score accounts for 100%, and the online evaluation list is updated in real time. The whole schedule is ranked at the end. The objective score after the reappraisal review is the final competition score.
Participation
During the event, a total of 1,588 teams participated, with 2,567 applicants.
Huawei competition topic: Intelligent Beam Prediction and Model Migration
It attracts students from many top universities to form teams. For example, students from Tsinghua University, Peking University, Chinese University of Hong Kong (Shenzhen), Fudan University, Sun Yat-Sen University and other universities participated in the competition and actively submitted excellent works, which also attracted wide attention from the industry.
Researchers from well-known companies such as Samsung, Lenovo, Telecom, China Mobile, China Unicom, Ericsson, and ZTE have signed up to participate.
Huawei competition topic: Smart Distributed Radio Map Construction
It attracts students from many top universities to form teams. For example, students from Tsinghua University, The Chinese University of Hong Kong (Shenzhen), Fudan University, Sun Yat-sen University and other universities actively participated in the competition and submitted excellent works, which also attracted wide attention from the industry.
Researchers from well-known companies such as Samsung, Lenovo, Telecom, China Mobile, China Unicom, ZTE, DOCOMO, etc. have signed up to participate.
OPPO competition topic: Wireless AI design for small sample condition scene adaptation and online update requirements
A total of contestants from more than 120 universities and scientific research institutes at home and abroad participated in the competition. For example, students from the University of Chinese Academy of Sciences, Beijing University of Posts and Telecommunications, Xi'an Jiaotong University, Fudan University, Nanjing University and other universities participated in the competition, actively participated in the competition and submitted excellent works, and also received wide attention from the industry.
Researchers from China Mobile, China Unicom, China Telecom, Huawei, ZTE, Samsung, Tencent, Baidu, Alibaba and other well-known companies have signed up to participate.
winning team
Special guests present awards to Huawei's winning teams
| "Smart Beam Prediction and Model Migration" Sety Award List | |
| Award name | winning team |
| first prize |
ECR5G |
| second prize |
X、Sorrymaker |
| third prize |
Fresh Clothes Angry Horse Boyhood, debias_world, Shangfen Squad |
| Winner |
Shmily, Intelligent Network Optimization, Horizon, Huangdao Cai Wenji, radio8212 |
Special guests present awards to Huawei's winning teams
| "Smart Beam Prediction and Model Migration" contest winner list |
|
| Award name |
winning team |
| first prize |
FDU-ssc |
| second prize |
hollynb, brother Shen eat garlic with coffee |
| third prize |
Shrek, UESTC-CINC, Chong Chong Chong |
| Winner |
REAL, FDU_Smart Science and Technology, Not So Much Progress, Visual Feast, Asphyxia |
Special guests present awards to the OPPO winning teams
| "Wireless AI Design for Scene Adaptation and Online Update Requirements for Small Sample Conditions" Contest Winner List |
|
| Award name |
winning team |
| first prize |
sunshine Boy |
| second prize |
Where does the AI go, may there be no Yasuo in heaven |
| third prize |
ECR5G, Sealed Exodia, Goose Mountain Swordsman |
| Winner |
Shell_Zs, LotusLink, today only learn Re-ID, don’t submit, just play~ |
Expert Reviews
Huawei competition topic: Intelligent Beam Prediction and Model Migration
Wireless Massive-MIMO technology has significantly improved the capacity of communication systems, and in high-frequency scenarios, more accurate and lower-cost intelligent beamforming is the key to achieving a greater increase in the capacity of future 6G communication systems. By analyzing the problems in beam prediction and migration, the contestants designed a variety of effective algorithms to achieve accurate beam prediction with a given measurement cost.
Starting from research points such as data enhancement, model design, model migration, and model training methods , the contestants analyzed environmental information and beam data, adopted machine learning or designed neural network models, and considered different training methods and migration methods to achieve accurate beam prediction .
The technical solutions adopted by the contestants basically cover the mainstream neural network model structure and machine learning algorithms in the existing AI technology , such as transformer, MAE, xgboost, etc. At the same time, according to the requirements of the competition, the model prediction accuracy index and the model generalization ability index were combined, and the average beam prediction accuracy under the three tasks (beam prediction, carrier frequency migration, and environment migration) reached above 0.9.
Huawei competition topic: Smart Distributed Radio Map Construction
The concept of radio maps is a key tool for wireless communication systems to meet 6G potential scenarios such as smart cities and smart factories. Explore the key technologies of 6G intelligent network through distributed learning method. The contestants analyzed the problems in the radio map, designed a variety of excellent algorithm solutions, and completed the construction of an intelligent distributed radio map.
Starting from research points such as data enhancement, model design, user scheduling algorithm, neural network structure, model fusion algorithm, and sparsification algorithm when uploading and distributing models, the contestants design user scheduling algorithms and neural network models by analyzing the environment and data, and consider the uplink and downlink costs at the same time to complete the construction of the radio map.
While effectively controlling the impact of communication overhead (uplink communication overhead and downlink communication overhead) in the wireless distributed learning system , the scores of the trained models (normalized mean square error of real signal strength and algorithm estimated signal strength) all reached more than 20 points. And through sparse algorithms such as dynamic quantization and pruning dropout, the communication overhead of uplink/downlink is greatly reduced.
OPPO competition topic: Scene adaptation for small sample conditions
and wireless AI design for online update requirements
Scenario-based and data-driven wireless AI solutions, while gaining high-performance advantages, also face the problem of generalization in multiple scenarios. For different scenarios, they need to bear high data collection costs and long training time to update the adaptation model. This poses a challenge to the deployment of AI-based solutions. Aiming at the topic of "Wireless AI Design for Scene Adaptation and Online Update Requirements for Small Sample Conditions" , this paper explores the construction method of wireless AI solutions under the limitation of small data in more practical target scenarios.
From the perspectives of data preprocessing/enhancement, transfer learning/meta-learning, model complexity control, and online adaptation model selection, the contestants realized a series of effective online update and scene adaptation solutions for wireless AI solutions under the limitation of small data through the optimization of data, learning, model, application and other stages.
Du Ying made a concluding speech for the competition
Du Ying, deputy director of the Mobile Communication Innovation Center of China Academy of Information and Communications Technology, pointed out that the participating teams and players actively contributed their wisdom, presented diversified solutions, and constantly refreshed the performance and efficiency of the competition questions, which deeply reflected the advantages of cross-integration technologies. These experiences and achievements will be applied to the follow-up 6G research work to jointly promote the integrated development of 6G and AI.
Du Ying said that the academic research and industrial practice of 6G and AI integration are getting deeper and deeper, and the organizers will continue to work hard, gather wisdom and practice, and jointly promote 6G to become an efficient digital base for intelligent and other applications.
During the competition, the DataFountain big data competition platform (referred to as DF platform) provided full-process operation services and technical support for this competition , continued to help the development of professional talents in technical fields such as 6G and AI integration, and promoted the benign construction of the industry ecology.
In the future, the DF platform looks forward to joining hands with more governments, enterprises, universities and research institutes to carry out more competitions and explorations related to "big data +" and "artificial intelligence +", select top digital intelligence technologies and talents at home and abroad, and add further assistance to my country's digital construction and digital economy development.