InTheLoop | 05.28.2013
CRD Researchers in the News: Wehner on Climate, Saye and Sethian on Bubbles
Following last week’s devastating tornado in Oklahoma, NBC News looked at the scientific questions behind apparent tornado patterns in “Curse or coincidence? Scientists study Tornado Alley's past and future.” For the question “Will climate change make tornadoes worse? More frequent?” they turned to Michael Wehner of the Computational Research Division, an expert on climate change and extreme weather. Wehner explains why those questions are difficult to answer now, but expresses optimism that improving supercomputer technology will make them answerable in the near future. Other media that picked up Wehner’s comments include The Independent (UK), National Geographic, Mother Nature Network, and the Huffington Post.
Robert Saye and James Sethian’s research on the mathematics of popping bubbles in a foam is drawing attention for its coolness factor as well as its practical applications. It’s the topic of articles in The Atlantic, New Scientist, Scientific American, Live Science, Time Magazine and HPCwire, among others.
Volunteer Opportunity: Oakland Non-Profit Seeking Software Engineers
The Women’s Initiative for Self Employment is seeking software engineer volunteers to work on extending the system that supports its clients. The Women's Initiative for Self Employment provides high-potential, low-income women with the training, funding and ongoing support to start their own businesses and become financially self sufficient.
To volunteer or learn more, write to firstname.lastname@example.org.
This Week’s Computing Sciences Seminars
Pseudo-Likelihood Methods for Robust Graphical Model Selection
Tuesday, May 28, 10:30–11:30 am, 50B-4205
Sang-Yun Oh, Simons Institute Postdoctoral Fellow, Stanford University
- Sparse high dimensional graphical model selection is a topic of much interest in modern day statistics. A popular approach is to apply l1 penalties to either (1) parametric likelihoods, or, (2) regularized regression/pseudo-likelihoods, with the latter having the distinct advantage that they do not explicitly assume Gaussianity. As none of the popular methods proposed for solving pseudo-likelihood based objective functions have provable convergence guarantees, it is not clear if corresponding estimators exist or are even computable, or if they actually yield correct partial correlation graphs. We proposes a new pseudo-likelihood based graphical model selection method that aims to overcome some of the shortcomings of current methods, but at the same time retain all their respective strengths. In particular, we introduce a novel framework that leads to a convex formulation of the partial covariance regression graph problem, which is then solved via a coordinate-wise approach with provable convergence guarantees. These guarantees ensure that estimators are well-defined under very general conditions, and are always computable. In addition, the approach yields estimators that have good large sample properties and also respect symmetry. Furthermore, application to simulated/real data, timing comparisons and numerical convergence is demonstrated. We also present a novel unifying framework that places all graphical pseudo-likelihood methods as special cases of a more general formulation, leading to significant insights; including equivalence between three of the four popular methods in the literature.
Link of the Week: Quantum Network Transmits Secure Video in Real Time
It's not quite a quantum internet — yet. But researchers at Los Alamos National Laboratory in New Mexico have developed a new, ultra-secure computer network that is capable of transmitting data that has been encrypted by quantum physics, including video files. The network, which currently consists of a main server and three client machines, has been running continuously in Los Alamos for the past two and a half years, the researchers reported in a paper released earlier this month. During that time, they have also successfully tested sending critical information used by power companies on the status of the electrical grid. Eventually they hope to use it to test offline quantum communication capabilities on smartphones and tablets. Read more.