On Supported Talents of 3rd CAAI Youth Talents Support Program
CAAI focuses on development of young AI talents. In 2017 CAAI formulated Implementation Rules of CAAI 2017-2019 Youth Talents Supporting Program and stipulated approaches of project establishment, talent recommendation and selection, and financial management. The Rules go accord with Management Approaches of CAST Youth Talents Supporting Program and Implementation Rules of CAST Youth Talents Supporting Program printed by China Association for Science and Technology (CAST). The rules aim to ensure smooth operation and financial security and interest of this program.
CAAI, as a gathering place and propeller for China AI youth talents, always prioritizes youth talent service and continuously explores selection and development mechanism of innovative S&T youth talents. CAAI helps youth talents realize their ideals and provides China with needed talented personnel to build it into a world-class S&T power.
Three talents are chosen for this program, and they are: Tianlei ZHANG from Beijing Trunk Tech Ltd., Ke GU from University of Science and Technology Beijing (USTB), and Lijun ZHANG from Nanjing University (NJU). All of them enjoy funding from CAST.
Associate Professor of Computer Science Department Nanjing University, Doctoral Supervisor
ZHANG specializes in large-scale machine learning and optimization, and is the author of more than 80 papers published in international academic conferences and journals, including over 50 CCF-A journal/conference papers. Many a time, ZHANG was invited to deliver special reports in important domestic academic conferences. ZHANG is also a senior director of many programs, including: General Programs/Youth Foundation of National Natural Science Foundation of China (NSFC), Youth Foundation of Jiangsu Provincial Natural Science Foundation, joint research program of Microsoft Research Asia (MSRA), applied research projects of Huawei Technologies Ltd. and Beijing Youku Technologies Ltd..
Assisted by the Program, ZHANG began basic and applied researches:
2020, General Programs of NSFC, Algorithms and Theories of Adaptive Online Learning.
2019, joint research program of MSRA, General Dynamic Regret for Online Learning.
2019, Beijing Youku Technologies Ltd., Online Learning for Recommendation System.
2019, Huawei Technologies Ltd., Research on Accelerated Optimization Technology Under Various Hardware Constraints.
In addition, ZHANG also participated in research projects, including: national key R&D projects, NSFC international (regional) cooperation and exchange projects.
With support from the Program, ZHANG has applied for and authorized 1 invention patent. Over 18 academic papers have been published, including: 8 on ICML (top-notch machine learning conference), NIPS/NeurlPS and COLT, 1 on JMLR (first-class machine learning journal), 12 on CCF-A conferences and journals. In 2018, ZHANG was elected Doctoral Supervisor of NJU.
Zhang is the CEO of TrunkTech, the secretary general of Intelligent Driving Professional Committee of Chinese Association for Artificial Intelligence (CAAI), and a member of Beijing Autonomous Driving Vehicles Road Test Expert Committee. He undertook the subject of “Autopilot Special Lane Design and Truck Queue Control” which belonged to the national key R&D program and the special project combining comprehensive transmission and intelligent transportation. He was also responsible for three major project supported by the National Natural Science Foundation of China. Among them the representative project was "The Research on Cloud Computing-based Key Technology of Mass Data Mining". Zhang participated in many provincial projects in Tianjin Port, Shanghai Lingang New Area, and Technology-supported Winter Olympics, etc. He had more than 10 years of research experience in the field of autopilot and had participated in the development of more than 10 types of autonomous driving vehicles. He published more than 60 patents and over 10 academic papers related to autopilot. Among those papers four were included in SCI\EI journals.
On April 12, 2018, the world's first autopilot electric collection truck went into experimental operations. After more than a year of actual operational testings, the Tianjin Port Group officially purchased four autopilot electric trucks through bidding in April 2019. In the following October, 21 additional orders were added, which have now been formally delivered and put into 24-hour actual ship operation. It took the lead in achieving a major breakthrough in large-scale applications. It was no doubt the world’s first automatic truck cluster working at the port.
In terms of artificial intelligence trucks for high-speed transportation, he jointly applied for the first road test license of automatic driving commercial vehicle in China on October 12, 2018.On May 7, 2019, a high-speed artificial intelligence truck queue was unveiled, launched and completed the test with excellent performance in China’s first large-scale car-following standard public verification test for commercial vehicles, and was recognized by the peer experts from China Automotive Technology and Research Center (CATARC) and the United Nations Standards Committee. At the end of 2019, Zhang’s team work was approved by the Ministry of Science and Technology as the “Major Special Project for Comprehensive Transmission and Intelligent Transportation”. They also planned to conduct large-scale road tests cooperating with many high-speed and logistics groups.
Gu is the professor and doctoral tutor of Beijing University of Technology, and he is also one of the Beijing’s specially appointed experts. He won the one and only Annual Best Paper Award of the top journal in the international multimedia field, IEEE TMM, as the first author of the paper. He also ranked tenth in the Automation and Artificial Intelligence Innovation Team Award of the Chinese Association of Automation, and was awarded the Outstanding Doctoral Dissertation of the Chinese Institute of Electronics. Since being selected in the Young Talent Supporting Program of the China Association for Science and Technology (CAST) in 2017, he had conducted many studies in the field of image quality evaluation and environmental quality perception. He accomplished a series of original works especially in five aspects, namely, the Internet screen image quality evaluation, virtual reality image quality evaluation, atmospheric fine particle estimation, smoke detection in petrochemical production and water turbidity monitoring of rivers. A total of 47 papers of Gu were published on IEEE Transactions, including 15 as first author and 12 as corresponding author during his time in the supporting program. Among those 47 papers, 9 were selected as ESI hot articles or ESI highly cited papers.
Gu presided over or participated in 5 scientific research projects, including hosting the Youth Project supported by the National Natural Science Foundation of China and the youth top-notch talent project based on Beijing’s High-level Innovation and Entrepreneurship Support Plan Youth Top Talent Project, and participating in the National Science and Technology Major Project, the National Key R&D Plan, and the Major Program of the National Natural Science Foundation of China.
Gu applied for more than 10 invention patents during the supporting period, including efficient image-based methods of PM2.5 concentration anticipation, a smoke detection method based on deep hybrid neural network, a Robust smoke detection method based on integrated deep convolution neural network, an air pollutant concentration monitoring method based on time weighting, a soft sensor for fine air particular PM2.5 based on image features and integrated neural network, reference-free image quality perception methods based on multi-scale natural scene statistics, air quality prediction methods based on iterative learning, reference-free image quality evaluation methods based on multi-scale analysis, and 3D synthetic image quality evaluation methods based on autoregressive local image description, etc.