Virtual seminar on How to Train Your Emu

Published on

Our Liverpool Virtual Seminar Series on Data Intensive Science will continue on Tuesday 10th January 2023 at 4pm GMT. The seminar will be given by Dr Kelly Moran, a research staff member in the statistical sciences group at Los Alamos National Laboratory who will present “How to train your Emu”.

Seminars in this series cover R&D outside of the data intensive science CDT’s core research areas and give an insight into cutting edge research in this area. At the end of the talk there will be a Q&A session with the speaker.

About the talk

An emulator is a fast statistical approximation to a complex simulation model. An emulator takes the same inputs as the simulation model and then predicts the output of the simulation. It trades off some accuracy for speed and can be used in time-sensitive applications like data analysis pipelines. In this talk, I'll describe the construction and some extensions of the Cosmic Emu, an emulator the matter power spectra with high accuracy. Noisy power spectra from simulations of varying resolutions are combined using a process convolution model to produce smooth estimates of power spectra for varying cosmological parameters. These are then combined using principal component decomposition and Gaussian process regression to construct an emulator. Time permitting, we'll discuss an approach for including partial simulation results, and other uses of Gaussian processes in the field of toxicology and epidemiology.

About the speaker

Dr. Kelly Moran began her career at Los Alamos National Laboratory (LANL) as a post-baccalaureate student in 2015. She did her PhD in Statistics at Duke University with the support of the Department of Energy's Computational Science Graduate Fellowship Program, returning to LANL as a summer student each year for projects with the Statistical Sciences group. She joined the group as a staff member after completing her PhD in 2020.

She has experience with jointly modelling related data streams, developing novel dimension reduction approaches, and accelerating computation for Bayesian methods. She has worked at the cutting edge of the field of disease modelling and forecasting, including contributing to the leading flu forecasting model in the U.S. and modelling verbal autopsy data in developing countries. Her work currently focuses on building statistical models for cosmology and heliospheric science applications. She is enthusiastic about helping students find a place to land for long-term careers at LANL, and particularly works to increase the presence of women and LGBT+ students in her division. She is also enthusiastic about the landscape of northern New Mexico, spending time outdoors, swing and tango dancing, and her cats.

How to attend

Participation is free, but you need to register to attend this and other webinars in the series. For more information and how to register please follow this link. Once registered, you will receive the Zoom connection details on the morning of the online seminar.

The seminar details

Speaker: Dr Kelly Moran (LANL)

Seminar title: “How to train your Emu”

Date/Time: Tuesday 10th January at 4pm GMT