On the afternoon of October 18th, the final day of the Zhejiang Lab Cup Global AI Competition kicks off in Hangzhou, welcoming 41 teams and 120 participants. Some participants have come from top-ranking universities around the world, such as Tsinghua University, Renmin University of China, Zhejiang University and the National University of Singapore; others have come from research institutions and enterprises such as the Institute of Computing Technology of the Chinese Academy of Sciences and Alibaba.
Since beginning three months ago the Zhejiang Lab Cup Global AI Competition has attracted worldwide attention from AI practitioners, teachers, students, and technophiles. More than 1,700 teams and 5,000 participants from all over the world signed up for the competition.
In the end, “RUC_AIM3” from Renmin University of China, “Ashen_one” from Zhejiang University, “Flying Fish” from Zhejiang Dahua Technology Co., Ltd. and “BugHunter” from Gridsum became the respective winners of the four tasks set.
A Gathering of AI Talents
Zhejiang Lab Cup Finals Now at a New Level
In recent years, artificial intelligence technologies, as they constantly reshape people’s lifestyles, have been on the fast track to becoming an important driving force behind a new round of technology revolutions and industrial reforms.
The Zhejiang Lab Cup Global AI Competition, which has been held annually for the past two years, has drawn the attention of some top-ranking teams around the world.
Some of these teams have come from renowned domestic universities and research institutes such as Peking University, Tsinghua University, Zhejiang University, Fudan University, and Renmin University of China; others have come from well-known internet enterprises such as Alibaba; and still others have come from some renowned international universities such as the Carnegie Mellon University of the United States, Technische Universität München of Germany, University of Sydney of Australia and Tokyo University of Japan.
“This year, we have better task design, better organization, better participation diversity, and a better competition atmosphere,” says Yuan Jixin, the general secretary of the organizing committee and deputy director of the lab.
Four Competition Tasks That Solve Real-World Problems
AI Application Accelerated
From AI defeating Lee Sedol to ever-evolving autopilot technology, and the growing number of dark factories, researchers are continuing to speed up the development of artificial intelligence technologies. In the future, there will be more application scenarios to be explored and more untapped areas to be found in industries such as education, retail, autopilot, and finance. Once impossible AI applications are becoming reality faster than ever.
The competition finals included two races: a technical challenge race and an innovation race. For these two races, there were four competition tasks: for the technical challenge race there was a video captioning task and for the innovation race there were three tasks: zero-shot object detection, multi-object tracking, and e-commerce review opinion mining.
“These problems are not just about basic theories. They also involve some of the latest achievements in artificial intelligence that can be put into use in the future,” says Zhu Shiqiang, the director of the lab. The competition is pushing forward the solutions to cutting-edge AI problems.
Moreover, the tasks of the competition have all been designed by researchers and insiders of the artificial intelligence industry, and the four tasks have shown a deep understanding of the artificial intelligence technologies and trends and are close to our everyday life.
“As e-commerce platforms continue to grow, when consumers are shopping online, they will count on previous reviews on a product by other consumers.” When an expert from the competition organization committee introduces the e-commerce review opinion mining task, he says, the task asks participants to analyze a tremendous amount of reviews and find information that is meaningful to consumers and merchants. The task “is inspired by and will be used to solve everyday problems.”
The video captioning task for the technical challenge race is a cross-media task that combines computer vision and natural language processing. The task requires AI models to recognize the objects shown in videos, understand their relationship, recognize their behaviors and express them in natural language.
“In the future, video captioning will play an important role in human-machine interaction, content-based video searching, and helping the visually impaired to understand the content of videos,” says an expert from the competition's organizing committee.
A High Ground for AI Innovation
Zhejiang Cup Now a Steppingstone for Talent Attraction
In recent years, the government of Zhejiang has made intensive and proactive planning efforts to catch up with global trends, adapt to national strategies and seize opportunities and, as a result, the government has made the development of artificial intelligence technologies a strategic priority. Against such a background, Zhejiang Lab has paid close attention to the latest developments of artificial intelligence technologies and taken proactive steps in planning its major artificial intelligence projects.
Last year, Zhejiang Lab launched the AI competition to make breakthroughs in fundamental research and core AI technologies. Meanwhile, the competition also aims to find AI talents and experts, inspire passion for artificial intelligence, and lay a solid foundation for building Zhejiang Lab as a high ground to speed up the development of artificial intelligence technologies.
The focus of the lab is currently AI 2.0. The lab is building a new generation open-source platform for artificial intelligence technologies and has carried out research in areas such as intelligent vision, brain-like computing, new machine learning methods and intelligent virtual training systems for autopilot. These efforts have been dedicated to pushing forward the frontier of artificial intelligence technologies and solving bottleneck problems in the field of artificial intelligence.
The 2019 Zhejiang Lab Cup Global AI Competition was co-organized by Zhejiang Lab, China Artificial Intelligence Industry Alliance, China Institute of Communications, and the municipal government of Hangzhou, under the guidance of the Cyberspace Administration of China and the provincial government of Zhejiang.