- You are here:
- GT Home
- Home
- News & Events

Series: Stochastics Seminar

Series: Stochastics Seminar

Series: Stochastics Seminar

Series: Stochastics Seminar

Series: Stochastics Seminar

Heavy tailed distributions have been shown to be
consistent with data in a variety of systems with multiple time
scales. Recently, increasing attention has appeared in different
phenomena related to climate. For example, correlated additive and
multiplicative (CAM) Gaussian noise, with infinite variance or heavy
tails in certain parameter regimes, has received increased attention in
the context of atmosphere and ocean dynamics. We discuss how CAM noise
can appear generically in many reduced models. Then we show how reduced
models for systems driven by fast linear CAM noise processes can be
connected with the stochastic averaging for multiple scales systems
driven by alpha-stable processes. We identify the conditions under
which the approximation of a CAM noise process is valid in the averaged
system, and illustrate methods using effectively equivalent fast,
infinite-variance processes. These applications motivate new
stochastic averaging results for systems with fast processes driven by
heavy-tailed noise. We develop these results for the case of
alpha-stable noise, and discuss open problems for identifying
appropriate heavy tailed distributions for these multiple scale systems.
This is joint work with Prof. Adam Monahan (U Victoria) and Dr. Will
Thompson (UBC/NMi Metrology and Gaming).

Series: Stochastics Seminar

Series: Stochastics Seminar

Series: Stochastics Seminar

Let (A_n) be a sequence of random matrices, such that for every n, A_n
is n by n with i.i.d. entries, and each entry is of the form b*x, where b
is a Bernoulli random variable with probability of success p_n, and x
is an independent random variable of unit variance. We show that, as
long as n*p_n converges to infinity, the appropriately rescaled spectral
distribution of A_n converges to the uniform measure on the unit disc
of complex plane. Based on joint work with Mark Rudelson.

Series: Stochastics Seminar

The seminar will be the third lecture of the TRIAD Distinguished Lecture Series by Prof. Sara van de Geer. For further information please see http://math.gatech.edu/events/triad-distinguished-lecture-series-sara-van-de-geer-0

Series: Stochastics Seminar

This talk concerns the description and analysis of a variational framework for empirical risk minimization. In its most general form the framework concerns a two-stage estimation procedure in which (i) the trajectory of an observed (but unknown) dynamical system is fit to a trajectory from a known reference dynamical system by minimizing average per-state loss, and (ii) a parameter estimate is obtained from the initial state of the best fit reference trajectory. I will show that the empirical risk of the best fit trajectory converges almost surely to a constant that can be expressed in variational form as the minimal expected loss over dynamically invariant couplings (joinings) of the observed and reference systems. Moreover, the family of joinings minimizing the expected loss fully characterizes the asymptotic behavior of the estimated parameters. I will illustrate the breadth of the variational framework through applications to the well-studied problems of maximum likelihood estimation and non-linear regression, as well as the analysis of system identification from quantized trajectories subject to noise, a problem in which the models themselves exhibit dynamical behavior across time.