Department of Mathematics
The Florida State University

This Week in Mathematics

26 - 30 January 1998

Monday: 26 January 1998

Graduate Seminar, 1:30 p.m., 204B Love Building

[no seminar]

Knot Theory Seminar, 3:35 p.m., 104 Love Building

Ivo Dinov, Florida State University
Wavelets, Decision Theory and Image Analysis
In this talk we will try to combine knowledge from the areas of Wavelet Analysis, Decision Theory and Parameter Estimation to address problems arising in Image Analysis.
We will motivate the study by looking at the problem of quantifying the performance of different image-registration techniques. Following an approach by Donoho & Johnstone we use decision theoretical methods and least-squares estimates to develop a meaningful (in an "optimal estimator" sense) scheme for denoising, analysis and comparison of signals in wavelet space.
Along the way we will discuss Multi-Resolution Analyses, wavelets, the discrete wavelet transform and their properties.
We will propose a new approach for "soft-thresholding" on the wavelet coefficients, and will exploit its optimality characteristics (regular and asymptotic).
This is intended as a two-period presentation.

Tuesday: 27 January 1998

Wednesday: 28 January 1998

(Real) Analysis Seminar, 2:30 p.m., 102 Love Building

Working through Carlos Kenig's lecture notes on oscillatory integrals and nonlinear pde.

Complex/Symbolic Coffee, 3:15 p.m., 105 Love Building
Complex/Symbolic, 3:35 p.m., 102 Love Building

Jack Quine, Florida State University
Geometric Parameters for the (2,2,2,3) Genus 3 Curve Family

Thursday: 29 January 1998

Algebra Seminar, 2:00 p.m., 104 Love Building

Paolo Aluffi, Florida State University
Algorithms in Invariant Theory

Topology Tea Time, 3:00 p.m., 204 Love Building
Topology Seminar, 3:35 p.m., 104 Love Building

Aleksandar Poleksic, Florida State University
Translation Numbers in Quasi-Convex Groups

Friday: 30 January 1998

Colloquium Coffee, 3:00 p.m., 204 Love Building
Colloquium, 3:30 p.m., 101 Love Building

Tamar Schlick, New York University
Speedup Up Biomolecular Simulations
The timestep problem in biomolecular dynamics will be introduced, and an efficient approach for propagating biomolecular dynamics according to the Langevin equation presented. The method termed LN, for its origin in a Langevin/Normal Modes scheme, arose fortuitously upon analysis of the range of harmonic validity of our normal-mode scheme LIN. LN combines force linearization with force splitting techniques. Unlike the competitive multiple-timestepping (MTS) schemes today --- formulated to be symplectic and time-reversible --- LN merges the slow and fast forces via extrapolation rather than "impulses"; the Langevin heat bath prevents systematic energy drifts. This combination succeeds in achieving more significant speedups than these MTS methods which are limited by resonance artifacts to an outer timestep less than half the period of the fastest motion (around 4--5 fs for biomolecules). Significantly, the frequency of updating the slow forces extends to 48 fs or more, resulting in speedup factors for LN exceeding 10. The application of LN to biomolecular dynamics is well suited for configurational sampling, thermodynamic, and structural questions.

Scientific Computing Seminar, 4:30 p.m., 200 Love Building

I. Michael Navon, Florida State University
Optimal Control Approach for a Shockwave Discontinuous Flow

Seminars and colloquia at "that other" university [a.k.a. the University of Florida]

Go to twims past or Main Math Menu