Invited Speakers



Assoc. Prof. Rui Han
Beijing Institute of Technology, China

Dr. Rui Han is a special research professor and associate professor at the School of Computer Science & Technology, Beijing Institute of Technology (BIT). Before joining BIT in 2014, He received MSc with honor in 2010 from Tsinghua University, China, and obtained his PhD degree in 2014 from the Department of Computing, Imperial College London, UK. His research interests are cloud and edge computing, edge intelligence, and system optimization for highly parallel workloads (in particular big data analytics and deep learning applications). He has over 50 publications in these areas, including papers at ACM MobiCom, TC, TPDS, TKDE, TDSC, INFOCOM, ICDCS, ICPP, CCGrid, and CLOUD.

Speech Title: LegoDNN: Block-grained Scaling of Deep Neural Networks for Vision Applications at Edge
Abstract: Deep neural networks (DNNs) have become ubiquitous techniques for vision applications. Executing DNNs on resource constrained edge devices exacerbates the existing challenges of meeting stringent latency requirements while maximizing inference accuracy. The prior art explores the accuracy-resource tradeoff by scaling the model sizes, but often incur large accuracy losses because of three limitations: (i) limited scaling space due to high overheads; (ii) imprecise latency estimation under performance interferences; and (iii) no online network structure optimization. For such an issue, we propose a lightweight, block-grained scaling approach termed LegoDNN. It automatically extracts several common blocks from a DNN model and quickly transforms each block into several descendant ones, whose combination provides massive online scaling options. At run-time, LegoDNN predicts the model's inference latency by estimating each block's latency separately while optimally combining the descendant blocks to maximize accuracy under deadline constraint. We implement LegoDNN in six typical vision applications and extensively evaluate it against state-of-the-art layer removing and nested network techniques using 12 popular DNN models. LegoDNN is available at https://github.com/LINC-BIT/legodnn.




Prof. Zheng-Ming Gao
Jingchu University of Technology, China

Zhengming Gao is an associate professor at Jingchu University of Technology. He received his D.-Eng. degree in 2010. He now serves as a faculty member with School of computer engineering, Jingchu University of Technology, Member of the Youth Working Committee of the Chinese Association of Artificial Intelligence, Chairman of Jingmen Greenby Network Technology Co., Ltd. He has finished eight major national defense projects, one provincial natural research project, four City Hall level projects. He has published more than eighty papers, of which sixties of them having been indexed in SCI/EI, he also occupied more than 50 patents and 40 software copyrights, he has published six monographs by now. He is now the leader of the “Research team of machine learning and its applications of Jingchu university of technology”, chairman with an institute of intelligent information technology, Hubei Jingmen industrial technology research institute; chairman with an institute of intelligent computation technology, Jingchu university of technology. And he is now focusing on intelligent information technology and development.

Speech Title: Nature-inspired Algorithms and Their Improvements
Abstract: Along with the development of our modern science and technology, we human understand nature more thorough in details. And when we describe them in mathematics, the equations blasted in dimensionality, complexity, and most of them could no longer be solved analytically any more. The nature-inspired algorithms (NIAs) have been a hot spot to find the best solutions of the hard problems, either unimodal or multimodal, scalable or non-scalable, symmetric or non-symmetric. In this report, we would focus on the simple beginning of the NIA, which reflects the work of our team. The NIAs would be classified and three types of classification methods would be reported, and we would further talk about the improving methods which could improve the capability of the NIAs.




Assoc. Prof. Rasha Ismail
University of Hertfordshire-GAF, Egypt

Dr. Rasha Ismail holds a Ph.D with specialization in Business Administration/MIS/E-business from the University of the West of England, in 2010 and an MBA in MIS/E-payment from Arab Academy for Science and Technology and Maritime Transport, Egypt. Since August 1996 to September 2012, she worked in the career of teaching until she got promoted to Assistant Professor at the Arab Academy for Science and Technology and Maritime Transport (AASTMT), Egypt. Rasha is currently working at the University of Hertfordshire as an Associate Dean of School of Business and a Program Leader of Business Administration. She has published a number of research about business process modeling, the transition to e-business and process improvements in education and in industry. Rasha Ismail received an award of high rank publication (Q2) from AUM in 2019.

Speech Title: Technological Advancements in E-Health
Abstract: E-health is the junction between medical informatics, public health businesses, health services, and sharing of information over the Internet. It has the main components that must essentially be aligned with countries’ goals, be financed, be policy generated, include expertise, and have an electronic foundation and electronic systems. Meanwhile, the latest advancements in technology introduced innovative applications supporting e-health. For instance, IoT, smartphones, virtual reality, and machine learning are brief examples of such advancements. The collection of such technologies can be integrated and thus can produce high-quality decision-making and problem-solving algorithms. Moreover, improved data analytics can be generated. These contain enhanced symptom detection, disease diagnosis, self-monitoring, rehabilitation, surgeries, and remote patient monitoring among other services. Furthermore, the presence of technology applications is not solely sufficient to trace health issues, the co-existence of custom and expert devices is closing the gap between tracking diseases and treating them. Though e-health is merited, there are challenges to the privacy and security of patients’ data. Sharing data within the cloud and among external bodies, requires full attention to protect it from intruders and business sharing.