Speaker: Eric Lownes (University of Maryland) -

Abstract: TBA

]]>

Speaker: Tao Luo (City University of Hong Kong) -

Abstract: http://www.terpconnect.umd.edu/~lvrmr/2018-2019-S/Classes/RIT.shtml

]]>

Speaker: Christian Zickert (UMD) -

Abstract: TBA

]]>

Speaker: Julian Yarkony (Verisk Analytics) -

Abstract: We present a novel approach for multi-person pose estimation (MPPE) using implicit column generation and nested benders decomposition. We formulate MPPE as a set packing problem over the set of person hypothesis (poses) in an image where the set of poses is the power set of detections of body parts in the image. We model the quality of a pose as a function of its members as described by a tree structured deformable part model.

Since we cannot enumerate the set of poses we attack inference using implicit column generation where the pricing problem is structured as a dynamic program and dual optimal inequalities are easily computed. We exploit structure in the dynamic program to permit efficient inference using nested Benders decomposition. We demonstrate the effectiveness of our approach on the MPII human pose annotation benchmark data set.

]]>

Speaker: Yassine Boubendir (New Jersey Institute of Technology) -

Abstract: This talk is divided into two parts. The first one focuses on iterative algorithms obtained by applying non-overlapping domain decomposition methods. We will specifically describe the importance of the choice of the transmission conditions and discuss the convergence of these methods. In the second part, we explain some iterative algorithms in the context of multiple scattering configurations and discuss important issues related to the high frequency regime.

]]>

Speaker: Shilin Yu (Texas A&M) -

Abstract: The coadjoint orbit method/philosophy suggests that irreducible unitary representations of a Lie group can be constructed as quantization of coadjoint orbits of the group. I will propose a geometric way to understand orbit method using deformation quantization, in the case of noncompact real Lie groups. This is joint work with Conan Leung.

]]>

Speaker: Jacob Bedrossian - University of Maryland | Department of Mathematics

Abstract: The purpose of this work is to perform a mathematically rigorous study of Lagrangian chaos and passive scalar turbulence in incompressible fluid mechanics. We study the Lagrangian flow associated to velocity fields arising from various models of fluid mechanics subject to white-in-time, Sobolev-in-space stochastic forcing in a periodic box. We prove that if the forcing satisfies suitable non-degeneracy conditions, then these flows are chaotic in the sense that the top Lyapunov exponent is strictly positive. Our main results are for the 2D Navier-Stokes equations and the hyper-viscous regularized 3D Navier-Stokes equations (at arbitrary Reynolds number and hyper-viscosity parameters). For the passive scalar problem, we study statistically stationary solutions to the advection-diffusion equation driven by these velocities and subjected to random sources. The chaotic Lagrangian dynamics are used to prove a version of anomalous dissipation in the limit of vanishing diffusivity, which in turn, implies that the scalar satisfies Yaglom's 1949 law of passive scalar turbulence in over a suitable inertial range -- the constant flux law analogous to the Kolmogorov 4/5 law. To our knowledge, this work is the first to provide a complete mathematical proof of any such scaling law from fundamental equations of fluid mechanics. The work combines ideas from random dynamical systems (the Multiplicative Ergodic Theorem and an infinite dimensional variation of Furstenberg's Criterion) with elementary approximate control arguments and infinite-dimensional hypoellipticity via Malliavin calculus. Joint work with Alex Blumenthal and Sam Punshon-Smith.

]]>