Arts & Sciences Events
This calendar presented by
Arts & Sciences
[PAST EVENT] Mathematics Colloquium and EXTREEMS-QED Lecture: Kevin McGoff (Duke University)
November 7, 2014
2pm - 3pm
Abstract: Dynamical systems arise frequently as mathematical models of physical and biological systems that evolve over time. In this context, I will focus on the question of system inference; that is, how should one estimate the structure or parameters of a model using observed time-series data?
Under suitable hypotheses, recent results (joint with S. Mukherjee, A. Nobel, and N. Pillai) show that maximum likelihood estimation (MLE) yields consistent estimates of the model as the number of observations tends to infinity. However, in many applied settings, MLE is computationally intractable. In the specific setting of inference of gene regulatory networks from time-series gene expression data, I will discuss a computationally efficient estimation method recently developed in joint work with X. Guo, A. Deckard, A. Leman, C. Kelliher, S. Haase, and J. Harer. Although gene regulatory networks will be used as a primary example throughout the talk, no prior knowledge of them will be assumed.
Under suitable hypotheses, recent results (joint with S. Mukherjee, A. Nobel, and N. Pillai) show that maximum likelihood estimation (MLE) yields consistent estimates of the model as the number of observations tends to infinity. However, in many applied settings, MLE is computationally intractable. In the specific setting of inference of gene regulatory networks from time-series gene expression data, I will discuss a computationally efficient estimation method recently developed in joint work with X. Guo, A. Deckard, A. Leman, C. Kelliher, S. Haase, and J. Harer. Although gene regulatory networks will be used as a primary example throughout the talk, no prior knowledge of them will be assumed.
Contact
Sarah Day