
Bao Wang
Mon May 14, 2018
4:00 pm
First, I will present the Laplacian smoothing gradient descent proposed recently by Prof. Stan Osher. We show that when applied to a variety of machine learning models including softmax regression, convolutional neural nets, generative adversarial nets, and deep reinforcement learning, this very simple surrogate of gradient descent can...

Jay Newby
Thu Apr 26, 2018
10:00 am
The longstanding view in chemistry and biology is that highaffinity, tightbinding interactions are optimal for many essential functions, such as receptorligand interactions. Yet, an increasing number of biological systems are emerging that challenge this view, finding instead that lowaffinity, rapidly unbinding dynamics can be essential for...

Rongjie Lai
Mon Apr 16, 2018
4:00 pm
Analyzing and inferring the underlying global intrinsic structures of data from its local information are critical in many fields. In practice, coherent structures of data allow us to model data as low dimensional manifolds, represented as point clouds, in a possible high dimensional space. Different from image and signal processing which handle...

Amir Moradifam
Mon Apr 9, 2018
4:00 pm
We study the inverse problem of determining both the source of a wave and its speed inside a medium from measurements of the solution of the wave equation on the boundary. This problem arises in photoacoustic and thermoacoustic tomography, and has important applications in medical imaging. We prove that if $c^{2}$ is harmonic in $\omega \subset \...

Keith Promislow
Mon Mar 5, 2018
4:00 pm
The self organization of phospholipids into membranes is fundamental to the origin of life, allowing for protection of internal structure while necessitating machinery to open pores. Evolutionary pressure has designed a myriad of controls in the guise of surface proteins that adjust the entropy of the lipidsolvent interactions...

Zhaosong Lu
Fri Feb 16, 2018
11:00 am
In the first part of this talk, we study a convexconstrained nonsmooth DC program
in which the concave summand of the objective is an infimum of possibly infinitely many smooth
concave functions. We propose some algorithms by using nonmonotone linear search and extrapolation
techniques for possible acceleration for this problem, and analyze their...

Xuehai Huang
Mon Feb 12, 2018
4:00 pm
A framework to systematically decouple high order elliptic equations into combination of Poissontype and Stokestype equations is developed using the tools of differential complexes and Helmholtz decompositions. The key step is to systematically construct the underling commutative diagrams involving the complexes and Helmholtz decompositions...