Search Constraints
Number of results to display per page
Results for:
Year
2022
Remove constraint Year: 2022
« Previous |
11 - 17 of 17
|
Next »
Search Results
-
Video
Geospatial information science is a discipline that focuses on using geospatial information technology to understand people, places, nature and processes of the earth. IoT refers to Internet of things, the combination of sensors, software and other technologies to connect and exchange data with other devices and systems over the Internet. The era of IoT brings us opportunities and challenges for geospatial information science. In the keynote, five characteristics and three scientific issues of geo-spatial information science in the era of IoT are summarised.
Event date: 06/09/2022
Speaker: Prof. Daren Li
Moderator: Prof. Christopher Chao (Hong Kong Polytechnic University)
Panel members: Prof. Qingyan Chen, Prof. Qinhao Chen (Hong Kong Polytechnic University)
Hosted by: PolyU Academy for Interdisciplinary Research
- Subjects:
- Land Surveying and Geo-Informatics
- Keywords:
- Internet of things Geospatial data Spatial data mining
- Resource Type:
- Video
-
Video
Adaptive computation is of great importance in numerical simulations. The ideas for adaptive computations can be dated back to adaptive finite element methods in 1970s. In this talk, we shall first review some recent development for adaptive methods with some application. Then, we will propose a deep adaptive sampling method for solving PDEs where deep neural networks are utilized to approximate the solutions. In particular, we propose the failure informed PINNs (FI-PINNs), which can adaptively refine the training set with the goal of reducing the failure probability. Compared with the neural network approximation obtained with uniformly distributed collocation points, the proposed algorithms can significantly improve the accuracy, especially for low regularity and high-dimensional problems.
Event date: 18/10/2022
Speaker: Prof. Tao Tang (Beijing Normal University-Hong Kong Baptist University United International College)
Hosted by: Department of Applied Mathematics
- Subjects:
- Mathematics and Statistics
- Keywords:
- Sampling (Statistics) Differential equations Partial -- Numerical solutions Mathematical models Adaptive computing systems
- Resource Type:
- Video
-
Video
Convex Matrix Optimization (MOP) arises in a wide variety of applications. The last three decades have seen dramatic advances in the theory and practice of matrix optimization because of its extremely powerful modeling capability. In particular, semidefinite programming (SP) and its generalizations have been widely used to model problems in applications such as combinatorial and polynomial optimization, covariance matrix estimation, matrix completion and sensor network localization. The first part of the talk will describe the primal-dual interior-point methods (IPMs) implemented in SDPT3 for solving medium scale SP, followed by inexact IPMs (with linear systems solved by iterative solvers) for large scale SDP and discussions on their inherent limitations. The second part will present algorithmic advances for solving large scale SDP based on the proximal-point or augmented Lagrangian framework In particular, we describe the design and implementation of an augmented Lagrangian based method (called SDPNAL+) for solving SDP problems with large number of linear constraints. The last part of the talk will focus on recent advances on using a combination of local search methods and convex lifting to solve low-rank factorization models of SP problems.
Event date: 11/10/2022
Speaker: Prof. Kim-Chuan Toh (National University of Singapore)
Hosted by: Department of Applied Mathematics
- Subjects:
- Mathematics and Statistics
- Keywords:
- Convex programming Semidefinite programming
- Resource Type:
- Video
-
Video
We introduce a Dimension-Reduced Second-Order Method (DRSOM) for convex and nonconvex (unconstrained) optimization. Under a trust-region-like framework, our method preserves the convergence of the second-order method while using only Hessian-vector products in two directions. Moreover; the computational overhead remains comparable to the first-order such as the gradient descent method. We show that the method has a local super-linear convergence and a global convergence rate of 0(∈-3/2) to satisfy the first-order and second-order conditions under a commonly used approximated Hessian assumption. We further show that this assumption can be removed if we perform one step of the Krylov subspace method at the end of the algorithm, which makes DRSOM the first first-order-type algorithm to achieve this complexity bound. The applicability and performance of DRSOM are exhibited by various computational experiments in logistic regression, L2-Lp minimization, sensor network localization, neural network training, and policy optimization in reinforcement learning. For neural networks, our preliminary implementation seems to gain computational advantages in terms of training accuracy and iteration complexity over state-of-the-art first-order methods including SGD and ADAM. For policy optimization, our experiments show that DRSOM compares favorably with popular policy gradient methods in terms of the effectiveness and robustness.
Event date: 19/09/2022
Speaker: Prof. Yinyu Ye (Stanford University)
Hosted by: Department of Applied Mathematics
- Subjects:
- Mathematics and Statistics
- Keywords:
- Convex programming Nonconvex programming Mathematical optimization
- Resource Type:
- Video
-
Video
Machine learning can deliver unprecedented performance. Its application domain has expanded into safety-critical cyber-physical systems such as UAVs and self-driver cars. However, the safety assurance of vehicular control has two conditions: 1) an analytical model of system behaviors such as provable stability, and 2) the software safety certification process (e.g., DO 178C) requires that the software be simple enough so that software safety can be validated by a combination of model checking and near exhaustive testing. Although ML software, as is, does not meet these two safety requirements, the real-time physics model supervised ML architecture holds the promise to 1) meet the two safety requirements and 2) enable ML software to safely improve control performance and safely learn from its experience in real-time. This talk will review the structure of the proposed architecture and some methods to embed physics into ML-enabled CPS control.
Event Date: 12/05/2022
Speaker: Prof. Lui Sha (University of Illinois Urbana-Champaign)
Hosted by: Graduate School
- Subjects:
- Aeronautical and Aviation Engineering and Computing
- Keywords:
- Machine learning Vehicles Remotely piloted Computer software -- Reliability Drone aircraft
- Resource Type:
- Video
-
Video
Using UC Berkeley as an exemplar, Prof. Koshland gave us a distinguished lecture on ‘Lighting the Way with Interdisciplinary Research since 1868’. One of the hallmarks of UC Berkeley has always been engagement of its faculty and students in research and education that expand in cross disciplines, joining on multiple approaches to address major challenges facing the world today, which is also what we are seeking to do at PAIR of PolyU. Moreover, Prof. Koshland shared with us the ways in which individuals and institutions can engage in interdisciplinary and multi-disciplinary research and education and how they can be creatively intertwined.
Event Date: 22/4/2022
Speaker: Prof. Catherine P. Koshland (University of California, Berkeley)
Moderator: Prof. Christopher Chao (Hong Kong Polytechnic University)
Panel members: Prof. Xiang-dong Li, Prof. Yuguo Li (Hong Kong Polytechnic University)
Hosted by: PolyU Academy for Interdisciplinary Research
- Keywords:
- Interdisciplinary research Interdisciplinary approach in education
- Resource Type:
- Video
-
Video
This presentation will discuss the future of global universities within the evolving context of current international education systems. It will begin with an overview of the challenges posed to modern day universities by the breadth and diversity of international further education curricular between the ages of 16 and 18 and the accessibility of world leading universities to international students, in general. Having established potential global trends in further education, the implications for global universities and higher education will be discussed within the context of both teaching and research
Event Date: 17/05/2022
Speaker: Prof. David Cardwell (University of Cambridge)
Moderator: Prof. Christoper Chao (Hong Kong Polytechnic University)
Panel members: Prof. Wong Kwok-yin (Hong Kong Polytechnic University), Prof. Andrew Cohen (Hong Kong University of Science and Technology)
Hosted by: PolyU Academy for Interdisciplinary Research
- Keywords:
- International education Universities colleges -- Administration Education Higher
- Resource Type:
- Video