News&Events

Focus

BIT’s Team Wins Best Paper Runner-Up Award of KDD 2021, an A-level Conferences Approved by CCF

the Best Paper Runner-Up Award on the conference

Recently, ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 2021 (KDD), one of the A-level conferences approved by CCF, announced the Best Paper Award. The team of Prof. Guoren Wang and Prof. Chi Harold Liu from the School of Computer Science and Technology, Beijing Institute of Technology (BIT) won the Best Paper Runner-Up Award in the Applied Data Science Track, which is the first time for BIT to win the Best Paper Award in CCF-A conferences. KDD 2021 received 1541 valid submissions, among which 238 papers were accepted. This means the acceptance rate is only 15.4%. Three Best Paper Awards were granted, accounting for 0.2%. The first author, Hao Wang, is a 2017 undergraduate student in Software Engineering, supervised by Prof. Chi Harold Liu, and was awarded the 2021 Xu Teli Scholarship.

The winning paper is entitled Energy-Efficient 3D Vehicular Crowdsourcing For Disaster Response by Distributed Deep Reinforcement Learning. Fast and efficient access to environmental and life data is key to the successful disaster response. Vehicular crowdsourcing (VC) by a group of unmanned vehicles (UVs) like drones and unmanned ground vehicles to collect these data from Point-of-Interests e.g., possible survivor spots and fire sites, provides an efficient way to assist disaster rescue. This paper proposes DRL-DisasterVC(3D), a distributed deep reinforcement learning framework, with a repetitive experience replay (RER) to improve learning efficiency, and a clipped target network to increase learning stability. This mechanism also uses a 3D convolutional neural network (3D CNN) with multi-head-relational attention (MHRA) for spatial modeling, and adds auxiliary pixel control (PC) for spatial exploration. The team designed a novel disaster response simulator, called “DisasterSim”, and conducts extensive experiments to show that DRL-DisasterVC(3D) outperforms all five baselines in terms of energy efficiency.

The VC system by a group of UVs for disaster rescue implemented in the paper

Professor Chi Harold Liu’s team has long been engaged in research on mobile group perception and unmanned edge group intelligence, with the goal of realizing a long time, large range, high quality and deep level information perception of mobile groups. The team has made progress in three aspects: perception space construction, task scheduling and data acquisition, which has been praised as “representative” by journals such as ToSN and IoTJ. Their achievements solve three fundamental problems: energy saving, information quality providing, and user participation stimulating. The research has developed a simulation platform for emergency management of drones, which was evaluated by the Seventh Academy of China Aerospace Science and Technology Corporation as “effectively supporting the overall demonstration and in-depth research of multi-sensor cooperative sensing technology in complex field environment”. The team received the Second Prize of Natural Science of the Chinese Institute of Electronics and China Industry-University Research Promotion Award.

KDD, one of the A-level conferences approved by CCF

Started in 1989, the KDD conference is the oldest and largest international top academic conference in the field of data mining, and the first conference to introduce the concepts of big data, data science, predictive analytics, and crowdsourcing. Google Scholar and Microsoft Academic have both ranked it as the number one international conference in the field of data mining. KDD 2021 received 1541 valid submissions, among which 238 papers were accepted. This means the acceptance rate is only 15.4%. Three Best Paper Awards were granted, accounting for 0.2%.

Prof. Chi Harold Liu

Chi Harold Liu is a professor and doctoral supervisor, a recipient of the National Outstanding Young Scientists Fund, the vice-president of the School of Computer Science and Technology of BIT, a fellow of the Chinese Institute of Electronics, a fellow of the Institution of Engineering and Technology (IET), and a fellow of the British Computer Society. He received his B.S. and Ph.D. degrees from Tsinghua University and Imperial College, UK, respectively. Liu has published 37 CCF-A papers, 6 ESI high citation papers, 22 domestic and foreign invention patents, with Google scholar citations of5141, and an H-index of 33. In addition, Liu was awarded the KDD 2021 Best Paper Runner-Up Award, the Second Prize of Natural Science of the Chinese Institute of Electronics, and the China Industry-University Research Promotion Award. He is currently a member of the Expert Advisory Group of the 14th Five-Year Plan of the National Information Industry, a member of the Technical Committee of the National Information Security Standardization Technical Committee, and an editorial board member of IEEE Transactions on Network Science and Engineering.

Website address:https://dl.acm.org/doi/pdf/10.1145/3447548.3467070

Baidu
map