One of the great, open challenges in machine vision is to train a
computer to "see people." A reliable solution opens up tremendous
possibilities, from automated persistent surveillance and
next-generation image search, to more intuitive computer interfaces.
It is difficult to analyze people, and objects in general, because
their appearance can vary due to a variety of "nuisance" factors
(including viewpoint, body pose, and clothing) and because real-world
images contain clutter. I will describe machine learning algorithms
that accomplish such tasks by encoding image statistics of the visual
world learned from large-scale training data. I will focus on
predictive models that produce rich, structured descriptions of images
and videos (How many people are present? What are they doing?) and
models that compensate for nuisance factors through the use of latent
variables. I will illustrate such approaches for the tasks of object
detection, people tracking, and activity recognition, producing
state-of-the-art systems as evidenced by recent benchmark
competitions.
We report on recent and ongoing work with Zhou Gang and I.M.
Sigal in which we prove that all MCF neckpinches are asymptotically
rotationally symmetric. Combined with recent work of other authors, this
represents strong evidence in favor of the conjecture that MCF solutions
originating from generic initial data are constrained to one of exactly
two asymptotic singularity profiles.
We prove a theorem of Woodin that, assuming $\mathsf{ZF} + \mathsf{AD}+ \theta_0 < \Theta$, every $\Pi^2_1$ set of reals has a semi-scale whose norms are ordinal-definable. The consequence of $\mathsf{AD}+\theta_0 < \Theta$ that we use is the existence of a countably complete fine measure on a certain set, which itself is a set of measures. If time permits, we outline how "semi-scale" can be improved to "scale" in the theorem using a technique of Jackson.
In the talk I will describe my recent work, joint with Carlos Kenig and
Fanghua Lin, on homogenization of the Green and Neumann functions for a family of second order elliptic systems with highly oscillatory periodic coefficients. We study the asymptotic behavior of the first derivatives of the Green and Neumann functions, using Dirichlet and Neumann correctors. As a result, we obtain asymptotic expansions of Poisson kernels and the Dirichlet-to-Neumann maps as well as optimal convergence rates in L^p and W^{1,p} for solutions with Dirichlet or Neumann boundary conditions.
In the talk, we will explain some joint work with Ovidiu Munteanu
concerning the geometry and analysis of complete manifolds with
Bakry-Emery Ricci curvature bounded from below.
Which natural numbers occur as the area of a right triangle with three rational sides? This is a very old question and is still unsolved, although partial answers are known (for example, five is the smallest such natural number). In this talk we will discuss this problem and recent progress that has come about through its connections with elliptic curves and other important open questions in number theory.
When does a majority exist? How does the geometry of the political spectrum influence the outcome? What does mathematics have to say about how people behave? When mathematical objects have a social interpretation, the associated theorems have social applications. We give examples of situations where sets model preferences, and prove extensions of classical theorems on convex sets that can be used in the analysis of voting in "agreeable" societies. This talk also features research with undergraduates.
The construction, maintenance and disruption of tissues emerge from the interactions of cells with each other, the extracellular microenvironment that the cells create and their external boundary conditions. Our ability to make biomedically meaningful predictions at the organ or organism level is limited because of the difficulty of predicting the emergent properties of large ensembles of cells. A middle-out approach to model building starting from cell behaviors and combining subcellular molecular reaction kinetics models, the physical and mechanical behaviors of cells and the longer range effects of the extracellular environment, allows us to address such emergence. I will discuss CompuCell3D as a multi-scale, multi-cell modeling platform to study such emergent phenomena and to connect them to their physiological outcomes. I will illustrate two projects using CompuCell3D, the development and of blood vessels and its effect on the growth of a generic model solid tumor and Choroidal Neovascularization (CNV) in Age-Related Macular Degeneration (the most common cause of blindness among the elderly). Time permitting, I will also briefly discuss our proof-of-concept simulations of somatic evolution in solid tumors.
In this talk, we show an electronic structure study making use of algebraic theory. Density functional theory (DFT) has become a basic tool for the study of electronic structure of matter, in which the Hohenberg-Kohn theorem plays a fundamental role in the development of DFT. Unfortunately, the existing proofs are incomplete. In the first part of this talk, we present a rigorous proof for the Hohenberg-Kohn theorem for Coulomb type systems using the Fundamental Theorem of Algebra. Kohn-Sham equation, a nonlinear eigenvalue problem, is the most widely used DFT model. In the second part, after using group theory to divide an eigenvalue problem into some groups of smaller ones that can be solved independently and in a two-level manner, we apply the decomposition approach to electronic structure calculations of symmetric cluster systems, in which we solve successfully thousands of Kohn-Sham eigenpairs with millions of degree of freedoms.
In 1982 Calabi proposed studying gradient flow of the L^2 norm
of the scalar curvature (now called Calabi flow) as a tool for finding
canonical metrics within a given Kahler class. The main motivating
conjecture behind this flow (due to Calabi-Chen) asserts the smooth long
time existence of this flow with arbitrary initial data. By exploiting
aspects of the Mabuchi-Semmes-Donaldson metric on the space of Kahler
metrics I will construct a kind of weak solution to this flow, known as a
minimizing movement, which exists for all time.