2023-2024 Colloquia
Abstract:
The discriminant Δ=𝑏2−4𝑎𝑐, of a quadratic 𝑎𝑥2+𝑏𝑥+𝑐 is a crucial parameter that provides definite information about the nature and number of solutions to the equation 𝑎𝑥2+𝑏𝑥+𝑐=0. For polynomials of degree higher than two the computation of discriminant becomes more complex and less informative. Therefore, it is natural to look for quantities that are simpler in nature and easier to compute to get partial information about the nature and number of solutions to polynomial equations. For example the cubic equation 𝑎𝑥3+𝑏𝑥2+𝑐𝑥+𝑑=0 has complex (non-real) solutions if 𝑏2−3𝑎𝑐<0 but 𝑏2−3𝑎𝑐>0 alone is not sufficient to conclude that the cubic equation must have only real solutions. In this talk we will present a method to generalize the discriminants of quadratics to polynomials of degree 𝑛>2 to obtain such partial information. Based on these generalized discriminants of quadratics we will see, for example, why there cannot be any quartic with four real roots having the last three terms ⋯+2𝑥2−𝑥+1 . Another application we will show is how to graph 𝑥4+(2𝑦2−4)𝑥2+𝑦4+4𝑦2+3=0 without using Calculus.
NYC College of Technology (City Tech))
Abstract:
We will present one of our recent research works on applying deep learning in abnormal traffic pattern detection and leverage the latest edge computing tools to develop a portable terminal solution for the system deployment.
Abstract:
Optimal control problems aim to determine the optimal evolution of a dynamical system over some time interval, where the system’s behavior can be influenced by a trajectory of independent control variables. These problems are often composed of highly nonlinear algebraic and differential equations for which a closed-form solution does not exist. Over the past few decades, several efficient computational solution techniques have been developed. For many problems, the best approach is to use so-called direct transcription methods, which transcribe the problem onto a discrete time domain and then solve the resulting finite-dimensional optimization problem with well-established algorithms. While this strategy is extremely effective for many applications, the algorithms can be computationally expensive if the transcribed optimization problem remains very complicated. This talk will review some of the main features of direct transcription, and then focus on sequential convex programming techniques, in which the transcribed optimization problem is solved with a series of simpler convex optimization problems. This approach often works very well in practice, and it is well-suited to some common optimal control problem structures. We will see how these methods can be used to solve a lunar landing trajectory optimization problem. Finally, we will consider the practical considerations when trying to apply fast optimal control algorithms to real space missions.
Join Zoom Meeting
https://us06web.zoom.us/j/85266842272?pwd=QXowJo4np0CLov3r6TPpq0GeJ350Fe.1
Meeting ID: 852 6684 2272
Passcode: iVzEm6
---
One tap mobile
+16469313860,,85266842272#,,,,*449055# US
+19292056099,,85266842272#,,,,*449055# US (New York)
---
Dial by your location
• +1 646 931 3860 US
• +1 929 205 6099 US (New York)
• +1 301 715 8592 US (Washington DC)
• +1 305 224 1968 US
• +1 309 205 3325 US
• +1 312 626 6799 US (Chicago)
• +1 386 347 5053 US
• +1 507 473 4847 US
• +1 564 217 2000 US
• +1 669 444 9171 US
• +1 669 900 6833 US (San Jose)
• +1 689 278 1000 US
• +1 719 359 4580 US
• +1 253 205 0468 US
• +1 253 215 8782 US (Tacoma)
• +1 346 248 7799 US (Houston)
• +1 360 209 5623 US
Meeting ID: 852 6684 2272
Passcode: 449055
Find your local number: https://us06web.zoom.us/u/kbNMaHxeKd
Abstract:
In this project, we develop randomized model reduction methods for advection-diffusion problems with sharply discontinuous source terms. To study such problems, we must solve the advection-diffusion equation, a partial differential equation (PDE) used to model systems such as a liquid dye being dissolved in a flowing fluid, or the combination of heat conduction and convection through a medium. This PDE arises often in the sciences, from fluid dynamics [1] to semiconductor physics [2], although in such contexts, the equation is almost always unsolvable by hand, so we rely on computer algorithms to efficiently obtain accurate approximations to the solutions. Our goal is to obtain a faster result than with direct numerical simulations like finite difference or finite element schemes. We employ randomized methods from data science to allow for parallel-in-time computation [3] and generation of a reduced order model. Compared to direct simulations, the reduced model produces a solution space of much lower dimension, meaning the computational complexity is greatly reduced. Thus, simulations with the reduced model allow for a further speedup in computational runtime, with the added benefit of maintaining accuracy. As a novel contribution, we consider the case of sharp discontinuities in source functions, partition the time domain into overlapping subintervals with overlap around the discontinuities, and construct a reduced basis on each subinterval. This allows for the construction of a reduced solution that combines information from the reduced bases, rather than relying on a single reduced basis to capture information throughout the time domain. We present a test case to show that this approach can provide significant improvements in accuracy compared to the construction of only one reduced basis over the whole-time domain.
One-sentence summary: To solve a partial differential equation modeling, e.g., the dissolving of a liquid dye in a flowing fluid or the conduction and convection of heat, we employ randomized methods from data science to facilitate parallel-in-time computation and generate a reduced order model to reduce computational complexity.
References
[1] Bird, R. B., Stewart, W. E. & Lightfoot, E. N. (2007). Transport Phenomena (Revised second). John Wiley & Sons.
[2] New, O. (2004) Derivation and numerical approximation of the quantum drift diffusion model for semiconductors, Jour. Myan. Acini. Arts & Sc., Vol. II (Part Two), No. 5.
[3] Schleuß, J., Smetana, K., & ter Maat, L. (2022). Randomized quasi-optimal local approximation spaces in time. arXiv preprint arXiv:2203.06276. To appear in SIAM J. Sci. Comput., 2023
Abstract:
Missing data imputation is a well-studied problem. Many of the methods solving this problem were developed and used in medical research. But the application of these methods is seen in many other areas, e.g., in property insurance underwriting.
Apart from very basic approaches of using the mean, mode, or other statistics of the non-missing observations for imputations, there are two other main categories of methods. The first category is the iterative approach. It is based on the idea of estimating the conditional distribution of one feature using all other available features. In one iteration, a conditional distribution estimator is trained to predict the value of each feature. This process is repeated for many iterations until it converges. This approach has been studied extensively and one of the most well-known methods is Multiple Imputation by Chained Equations (MICE).
Lately another approach using deep generative models was developed. In this approach, a generative model is trained to generate values in missing parts based on observed values. Review of the methods and specifically for tabular data (numeric and categorical) will be followed with examples of my industry experience with them.
Abstract: