Mathematics Colloquium Schedule

Fall 2014

 

Coming ...

 

Date 9-12-2014 (12pm-1pm, F03-200A), Professor Ko Honda, UCLA

Title: An invitation to Floer homology

Abstract: This is a gentle introduction to Floer homology. ``Floer
homology'' is a generic term for various homology theories of knots, 3- and
4-dimensional manifolds (aka spaces), symplectic manifolds, contact
manifolds, etc., and has had an enormous impact in geometry/topology since
its introduction by Floer more than twenty years ago. In this talk we start
with a baby version of this theory called Morse homology, which gives a way
to distinguish topological spaces (e.g., a sphere from the surface of a
donut). We then build our way up to more recent theories such as contact
homology and embedded contact homology.

Date 9-19-2014 (12pm-1pm, F03-200A), Dr. Ryan Compton , HRL

Title: Geotagging One Hundred Million Twitter Accounts with Total Variation
Minimization

Abstract: Geographically annotated social media is extremely valuable for modern
information retrieval. However, when researchers can only access
publicly-visible data, one quickly finds that social media users
rarely publish location information. In this work, we provide a method
which can geolocate the overwhelming majority of active Twitter users,
independent of their location sharing preferences, using only
publicly-visible Twitter data.

Our method infers an unknown user's location by examining their
friend's locations. We frame the geotagging problem as an optimization
over a social network with a total variation-based objective and
provide a scalable and distributed algorithm for its solution.
Furthermore, we show how a robust estimate of the geographic
dispersion of each user's ego network can be used as a per-user
accuracy measure, allowing us to discard poor location inferences and
control the overall error of our approach.

Leave-many-out evaluation shows that our method is able to infer
location for 101,846,236 Twitter users at a median error of 6.33 km,
allowing us to geotag over 80% of public tweets.