Assco. Prof. Pavel Loskot
IEEE Senior Member
Zhejiang University-University of Illinois at Urbana-Champaign Institute (ZJUI), China

Areas of Expertise: Statistical signal processing, Probabilistic modeling, Networked systems

Bio: 25+ years of experience in design, analysis, implementation and deployment of telecommunication systems through numerous academic and industrial collaborative projects and consultancy contracts. Expert level knowledge of digital and statistical signal processing, algorithms and methods. Solid background in applied probability and statistics. Avid Linux programmer and user since 1996. In 2014/2015, as a Visiting Researcher at CSRC of the Chinese Academy of Engineering Physics started working on computational molecular biology. In 1999-2001, Research Scientist and Project Manager at CWC, Oulu, Finland. A Fellow of the Higher Education Academy of the UK, and the Recognised Research Supervisor of the UK Council for Graduate Education. A Senior Member of the IEEE since 2013.

Speech title: From Sensitivity Analysis to Bayesian Optimization

Speech abstract: A typical situation in designing data processing systems including the algorithms for signal processing and the models for machine learning is the presence of a large number of parameters. This constitutes a very challenging configuration problem in many dimensions how to choose the values of the parameters yielding the optimum or at least the desirable performance. For instance, the ablation experiments in machine learning assess the influence of each feature, but removing the features individually one by one. While this can provide at least some insight into the effect of each feature on the performance, it is well known that one-at-a-time (OTA) analysis can lead to misleading conclusions. In this talk, I will outline basic methods of sensitivity analysis from sensitivity indices, to different methods of decomposing the output variance, to creating surrogate models, and finally explain how the Bayesian optimization has been successfully applied to many high-dimensional optimization problems including finding the hyperparameters of machine learning models.

Prof. Xin Nie
Wuhan Institute of Technology, China

Areas of Expertise: Software engineering, intelligent optimization algorithms, machine vision, SoC Design

Speech title: Coordinated Task-planning for Multi-autonomous Satellites

Speech abstract: With the rapid development of remote sensing technology, multi-satellite autonomous task planning has become a key research direction in satellite Earth observation. The core challenge in this field lies in how to achieve efficient collaborative planning of Earth observation tasks under the conditions of limited resources and diversified observation requirements. Intelligent optimization algorithms, as an effective tool to solve this problem, are continuously promoting technological progress in this field. Firstly, an overview of the importance of multi-satellite collaborative task planning and its current research status at home and abroad will be presented. The observation task planning of satellite constellations not only needs to consider the observation capabilities of satellites, user requirements, and resource constraints but also has to deal with complex issues such as visibility constraints between stars and the Earth and spatial geometric transformations....more...