Mike Finegold
Research
I am interested in applying statistical methods to solve practical
real-world problems - almost all of my research is motivated this
way. These real-world problems come from consulting clients
(currently in legal cases and edcuational testing); from my
appointment with LARC (Living Analytics Research Centre); from other
departments at CMU (Six Degrees of Francis Bacon, for example); and
from the wide world beyond our doors (including a statistical method
for adaptive keyboard design).
While this leads to a variety of research projects, one common theme
is network inference. From different data types (gene expression
levels, digitized texts, chat messages, transactions histories,
etc.) what can we infer about the existence,
strength, and type of relationship between various entities (usually
people - but sometimes genes, for example - and usually living
people - but not always. Francis Bacon, for one, is dead).
Project Descriptions
Six Degrees of Francis Bacon is an interdisciplinary project, partially
funded by a Google Faculty Research Award, to
recreate the early modern social network in Britain. Very preliminary
results can be found on the project website, articles
are forthcoming, and I will be presenting at Digital Humanities 2013
in July.
I am currently visiting LARC (Living Analytics Research Centre), to
work on big data problems related to networks and experimental
design. I have worked on
relationship
strength in social networks; experimental design for measuring
the impact of interventions in social networks (one article in
process, another to be presented at ICWSM 2013); click fraud
analysis (under review) and ad targeting in mobile ad networks (long
way to go).
Previous work on statistical methodology for inference of gene
networks can be found on arXiv, the
Annals of Applied Statistics, and the Proceedings of the 25th Conference on Uncertainty in
Artificial Intelligence.