LIFE AT FIELDS
Statistics Graduate
Student Research Day
April 25, 2019 • Fields Institute
Organizers: Michaël Lalancette, Radu Craiu,
Kamal Ra, Jeffrey Rosenthal, Qiang Sun,
Stanislav Volgushev, and Zhou Zhou
(University of Toronto)
The Statistics Graduate Research Day is an
annual one-day conference organized by the
Department of Statistical Sciences (DOSS) at
the University of Toronto and the Statistics
Graduate Students Union. First organized in
2009, the event offers graduate students an
opportunity to learn from and be introduced
to leading experts, and to present their work
in either a talk or poster competition.
The 2019 edition of Research Day focused on
high-dimensional inference and computation, and
their applications to big data. The slate of speakers
at the event was of exceptionally high stature and
included a recipient of the 2018 COPSS award,
Richard Samworth, who presented the idea of
shape-constrained non-parametric
density estimation with an emphasis on
high‑dimensional log-concave densities.
Biostatistician Bhramar Mukherjee spoke of
the ongoing and unprecedented information
and data revolution and the potential impact
on the field, particularly for those who are
classically trained. The audience also heard
from Edward George, whose talk demonstrated
how the spike-and-slab methodology often
used in Bayesian variable selection can be used
for likelihood-based estimation and variable
selection in a linear regression framework.
The final speaker of the day, Elizabeth Schifano,
outlined recent work in the analysis of big stream
data, defined as large amounts of data being
received in streams, with a particular focus on
survival data. Developing methodology in the
18
framework of big data requires significant
thought to produce statistically sound
inference in a computationally feasible way.
Students from DOSS made presentations
that exposed the audience to the promising
statistical research done at the University
of Toronto on subjects including learning
theory, mathematical finance, Bayesian non-
parametric estimation, and survey sampling
theory. Graduate students also showcased
their research in a poster competition.
Feedback from the participants, including
the esteemed invited speakers, praised
the organization of the conference, the
quality of the research presented, and the
fruitful exchange between participants.