Resolution analysis in some scattering problems and super-resolution in certain scenarios
Dr. Yat Tin Chow from UC Riverside
September 18, 2020
12:00pm-1:00pm via Zoom
Join 9/18 Zoom meeting
Meeting ID: 910 8240 4817
In this talk, we explore image resolution and ill-posed-ness of inverse scattering problems. In particular, we would like to discuss how certain properties of the inclusion might induce high-resolution imaging. We first explore the super-resolution phenomenon with certain particular high contrast inclusion. We then discuss how local sensitivity (and resolution) around a point is related to the extrinsic curvature of the surface of inclusion around the point. Along the line, we also discuss concentration of plasmon resonance (in a certain manner) at boundary points of high curvature leveraging the Heisenberg picture of quantization and quantum ergodicity first derived by Shnirelman, Zelditch, Colin de Verdiere and Hellfer-Martinez-Robert. This is a joint work with Habib Ammari (ETH Zurich), Hungyu Liu (CityU of HK), Keji Liu (Shanghai Key Lab), Jun Zou (CUHK).
About the Presenter
Yat Tin Chow is currently an Assistant Professor in the Department of Mathematics. He received his Ph.D. in Mathematics from the Chinese University of Hong Kong. He joined the faculty in UC Riverside after being a CAM assistant adjunct professor in Department of Mathematics in UCLA. His major research direction is applied mathematics. Dr. Chow's current research interests includes resolution analysis and enhancement of imaging from boundary measurements of various physical quantities, e.g. electric current, acoustic wave, light intensity, etc. He is also interested in computational methods of medical imaging and tomography, e.g. Electrical Impedance Tomography. Dr. Chow's other fields of interest include both theoretical and numerical aspects of large scale optimization method, computations of control methods and conflict modeling in high dimensional systems, as well as transportation plans and games between large populations in the mean field, and different phenomena that arise from this setting. If you are interested in his search areas, kindly visit Dr. Chow's personal website.
The Mathematics Colloquium is a unique opportunity for students to learn about new developments in mathematics and what mathematics and statisticians do after they graduate. Hosted by the Department of Mathematics and Statistics at California State University, Long Beach, the weekly meetings invite guests from universities, research laboratories, and industry to present and discuss current topics in mathematics. All students are encouraged to attend.
Additional Colloquia will be listed here.
The Mathematics Colloquium Archive has the Colloquia from previous semesters.
8/28/20: Creativity-in-Progress Reflections (CPR) on Proving and Problem Solving
Dr. Gülden Karakök from University of Northern Colorado
Numerous reports, policy and standards documents, and research studies emphasize the importance of creativity. For example, the recent report from the World Economic Forum noted that creativity at work is one of the top-three demanded skills, and that it "has jumped from 10th place to third place in just five years" (Schöning & Witcomb, 2017, para. 12). Within the domain of mathematics, similar emphases are made by mathematicians, mathematics education researchers and policy/standards makers. For example, the Mathematical Association of America’s (MAA) CUPM Curriculum Guidelines (Schumacher & Siegel, 2015) for majors in the mathematical sciences states that "a successful major offers a program of courses to gradually and intentionally lead students from basic to advanced levels of critical and analytical thinking, while encouraging creativity and excitement about mathematics" (p. 9). In this talk, I will briefly summarize some of the research on mathematical creativity at the K-16 levels and introduce the work of the Creativity Research Group focusing on undergraduate mathematics courses. Our research group aims to explore ways in which undergraduate students' mathematical creativity can be fostered and explicitly valued in mathematics courses that include proof-construction and/or problem solving activities. I will introduce the Creativity-in-Progress Reflections (CPR) on Proving and Problem Solving tools that we designed. These formative assessment tools were created to enhance mathematical creativity (of users) while facilitating proof-construction and problem-solving heuristics as well as fostering metacognition. With two categories, Making Connections and Taking Risks, these formative assessment tools aim to develop mathematical discourse centered around aspects of creativity related to fluency, elaboration, flexibility, and originality. I will provide some examples of how one might implement these tools in various mathematics courses as well as discuss some illustrative empirical examples from our research studies.
9/4/20: Property/Casualty Insurance Ratemaking
Prof. Janet Duncan from FCAS, FSA, MAAA
Insurance is a contractual promise to reimburse policyholders for future losses. Consumers often comment that they don't understand their insurance rates – it all seems very mysterious to them. But in reality, creating insurance rates is very logical when broken down into its component parts. The fundamentals of insurance ratemaking are very similar to the pricing of many other products, i.e., understanding cost and determining a target profit load. The major difference is that for many products, the cost is easily determined from the manufacturing process. However, for insurance, the cost involves significant uncertainty about the future. This presentation will introduce the audience to fundamental insurance principles and the mystery behind insurance ratemaking.
About the Presenter
Professor Janet Duncan has over 30 years of property/casualty financial analysis experience, including commercial and personal lines reserving and pricing, financial and capital modeling, planning, and management reporting. Janet's work experience includes six years as CNA's senior vice president and signing actuary, responsible for $17 billion of property/casualty reserves, including standard commercial lines, specialty lines, and discontinued operations. Prior to CNA, Janet worked at XL Capital, serving in roles of increasing responsibility including executive vice president and chief finance officer of XL Insurance Europe and Asia. She also worked with PricewaterhouseCoopers LLP (consulting and audit support), and served in various actuarial roles with Aetna Life & Casualty where she began her insurance career. Janet has a bachelor's degree in Math/Actuarial Science from the University of Connecticut. She has served on many actuarial committees including the CAS Committee on Professionalism Education, the CAS Committee on Reserves, the AAA IFRS Task Force, the AAA Opinion Seminar Committee, and the SOA Strategic Planning Task Force. She is now working as a lecturer at the Department of Applied Probability and Statistics at UC, Santa Barbara.
9/11/20: Fast Graph-based Algorithms for Analyzing Protein-Protein Interaction Networks
Dr. Junyuan Joanne Lin from Loyola Marymount University
This research aims to predict proteins' functions from protein-protein interaction (PPI) networks. The PPI networks we study include physical and genetic interactions between labeled and unlabeled proteins. This allows us to predict proteins' unknown functions based on the function labels of closely interacted "neighbors". In this presentation, I will present our award-winning graph-based algorithms that achieve the best prediction accuracy worldwide in the 2016 Disease Module Identification DREAM Data Mining Challenge. We define the diffusion state distance (DSD) metric, which sets appropriate distances to measure proteins' proximity on PPI networks as well as many other close-knit networks including social and energy networks. Fast algorithms, such as the unsmoothed aggregation algebraic multigrid method with random projections, are adopted to compute the DSD efficiently. Based on random walks combining with random projections, we propose graph-based methods to construct k-nearest-neighbor (kNN) graphs under the DSD metric for function prediction. We test our proposed algorithms on different networks to demonstrate that the computational cost of the algorithms is nearly optimal.