Description
Speaker: Dr. Iker Perez, Assistant Professor in Statistics at the
University of Nottingham, UK
Date: 30th Aug 2018
Time 3-4 pm
Venue: Ramanujan Hall
Title: Approximate Uncertainty Quantification with Jump Processes
Abstract:
This talk will discuss foundational statistical challenges and
probabilistic considerations associated with families of stochastic jump
models, which often find applications in domains such as genetics,
epidemiology, mathematical biology or operations research. I will review
Markov jump processes, and by means of common accessible examples, discuss
the strong impediments posed by real-world application scenarios to inverse
uncertainty quantification tasks.
Next, I will discuss current statistical advances linked to structured jump
systems along with relevant literature. Through a model exemplar borrowed
from queueing theory, I will finally present an approximate and scalable
variational Bayesian framework, suitable for uncertainty quantification
tasks with a large class of structured processes. The talk will further
include examples with applications of the methods introduced.