Faculty Meeting
Questions or comments?

Posted April 21, 2014

3:00 pm Lockett 009Meeting of tenured faculty

Topology Seminar
Seminar website

Posted April 21, 2014

3:30 pm - 4:20 pm Lockett 233
Mustafa Hajij, Department of Mathematics, LSU
Graduate Student

Virtual Topology Seminar: " Skein Theory and q-Series"

Abstract: We study the tail a q-power series invariant of a sequence of admissible trivalent graphs with edges colored n or 2n. We use local skein relations to understand and compute the tail of these graphs. This allows us to understand the tail of the colored Jones polynomial for a large class of knots and links. For many quantum spin networks they turn out to be interesting number-theoretic q-series. In particular, certain quantum spin networks give a skein theoretic proof for the Andrews-Gordon identities for the two variable Ramanujan theta function as well to corresponding identities for the false theta function. Finally, we give product formula that the tail of such graphs satisfies.

Faculty Meeting
Questions or comments?

Posted February 26, 2014

3:00 pm Lockett 009Faculty meeting

Pasquale Porcelli Lecture Series
Special Lecture Series

Posted March 12, 2014

Last modified March 13, 2014

Susan Murphy, University of Michigan H.E. Robbins Professor of Statistics & Professor of Psychiatry, Research Professor, Institute for Social Research
2013 MacArthur Fellow

L 1: Getting SMART about Adapting Interventions L2: Adaptive Confidence Intervals for Non-smooth Parameters

Lecture 1 for General Audience (3:30-4:20)

Getting SMART about Adapting Interventions

Abstract: Imagine you are a child with ADHD. Wouldn't you like your

doctors to periodically adapt your treatment to your unique--and

ever-changing--condition? And wouldn't you be excited to learn that an

algorithm used to analyze your medical data was originally developed for

applications in robotics and artificial intelligence? This lecture will

explain how a randomized clinical trial design (Sequential Multiple

Assignment Randomized Trial or SMART) is being used to develop adaptive

interventions--protocols that systematize sequential decision-making that

is key to effective treatment of health problems. Examples include a

study of children with ADHD and an ongoing study to improve care at

mental health clinics.

Lecture 2 for more specialized audience (4:30-5:20)

Adaptive Confidence Intervals for Non-smooth Parameters

Abstract: Non-regular, aka "non-smooth" parameters are of scientific

interest occur frequently in modern day inference. In particular when

scientific

interest centers on a non-smooth function of regular parameters such as in

the assessment of a machine learning classifier's performance, in the

estimation of multistage decision making policies and in the use of

methods that use assumptions of sparsity to threshold estimators. If

confidence intervals are

considered at all, most research assumes potentially implausible

"margin-like" conditions in order to justify the proposed confidence

interval method. We describe a different approach based on

constructing smooth upper and lower bounds on the parameter and then

basing the confidence interval on the smooth upper and lower bounds.

In particular two settings will be discussed and contrasted, that of a

confidence interval for the mis-classification rate and a confidence

interval for a parameter in multistage decision making policies.

Applied Analysis Seminar
Questions or comments?

Posted March 26, 2014

Last modified April 21, 2014

Mark Wilde, LSU Department of Physics/CCT

Renyi generalizations of the conditional quantum mutual information

Abstract: The conditional quantum mutual information I(A;B|C) of a tripartite quantum state on systems ABC is an information quantity which lies at the center of many problems in quantum information theory. Three of its main properties are that it is non-negative for any tripartite state, that it decreases under local operations applied to systems A and B, and that it obeys the duality relation I(A;B|C)=I(A;B|D) for a four-party pure state on systems ABCD. It has been an open question to find Renyi generalizations of the conditional mutual information, that would allow for a deeper understanding of the original quantity and find applications beyond the traditional memoryless setting of quantum information theory. The present paper addresses this question, by defining different Renyi generalizations of the conditional mutual information that all converge to the conditional mutual information in a limit. Furthermore, we prove that many of these generalizations satisfy the aforementioned properties. As such, the quantities defined here should find applications in quantum information theory and perhaps even in other areas of physics, but we leave this for future work. We also state a conjecture regarding the monotonicity of the Renyi conditional mutual informations defined here with respect to the Renyi parameter. We prove that this conjecture is true in some special cases and when the Renyi parameter is in a neighborhood of one. Finally, we discuss how our approach for conditional mutual information can be extended to give Renyi generalizations of an arbitrary linear combination of von Neumann entropies, particular examples including the multipartite information and the topological entanglement entropy. This is joint work with Mario Berta (Caltech) and Kaushik Seshadreesan (LSU). This is based on the recent paper http://arxiv.org/abs/1403.6102

Actuarial Student Association Meeting

Posted March 21, 2014

Last modified April 22, 2014

Actuarial Student Association sponsored visitor

Our guest, Shelley Johnson (Foster & Foster Inc), is the consulting actuary to both the Louisiana State Employees' Retirement System (LASERS) and the Teachers Retirement System of Louisiana (TRSL).