banner USF Home College of Arts & Sciences OASIS myUSF USF A-Z Index

USF Home > College of Arts and Sciences > Department of Mathematics & Statistics

Mathematics & Statistics
pendulum
  Colloquium Archive

Colloquia — Fall 2016

Friday, October 28, 2016

Title
Speaker

Time
Place
Sponsor

TBA
Peter Ebenfelt
University of California, San Diego
3:00pm-4:00pm
CMC 130
Dima Khavinson









Abstract

TBA

Friday, October 7, 2016

Title
Speaker

Time
Place
Sponsor

Phase Transitions in Community Detection
Cris Moore
Santa Fe Institute
3:00pm-4:00pm
CMC 130
Razvan Teodorescu

Abstract

There is a deep analogy between statistical inference and statistical physics. I will give a friendly introduction to both of these fields, describing how the posterior distribution of a model given data is treated as the Boltzmann distribution of an appropriate physical system. I will then discuss phase transitions in community detection in networks, and clustering of sparse high-dimensional data, where if our data becomes too sparse or too noisy it suddenly becomes impossible to find the underlying pattern, or even tell if there is one. Along the way, I will visit ideas from computational complexity, random graphs, random matrices, and spin glass theory.

Friday, September 30, 2016

Title
Speaker
Time
Place

Structured Decision-Making with Multiple Objectives
Lu Lu
3:00pm-4:00pm
CMC 130

Abstract

With increasingly constrained budgets, it is desirable to extract more information with limited data collection resources and make key decisions by balancing multiple objectives. A structured decision-making process using the Pareto front approach is presented to facilitate informed decision-making based on considering all possible solutions and trade-offs as well as understanding potential impacts from subjective choices. The method uses a two-stage process, where the first objective stage identifies the collection of all superior choices on the Pareto Front with an optimization search, and the second stage uses a rich set of graphical tools to quantitatively assess the trade-offs and robustness of different solutions to user priorities captured in weighting, scaling, and metric form choices. The final solution is a justifiable choice tailored to match what is important to the user. Applications in cost-quality choices and design of experiments serve as illustrating examples. Other applications in reliability analysis, resource allocation, as well as multiple response optimization are explored as well. The methodology is very general and can be applied to many applications with flexible choices of metrics for quantifying objectives of interest.