[Colloquium] Tchoua/MS Presentation/Feb 4, 2016

[Margaret Jaffey]

This is an announcement of Roselyne Tchoua's MS Presentation.

------------------------------------------------------------------------------
Date:  Thursday, February 4, 2016

Time:  9:00 AM

Place:  Ryerson 276

M.S. Candidate:  Roselyne Tchoua

M.S. Paper Title: HYBRID HUMAN-COMPUTER APPROACH TO SCIENTIFIC FACTS
EXTRACTION FROM LITERATURE: AN APPLICATION IN MOLECULAR ENGINEERING

Abstract:
A wealth of valuable research data is locked within the millions of
research articles published every year. Reading and extracting
pertinent information from those articles has become an unmanageable
task for scientists. This problem hinders the advancement of science,
making it hard to build on existing results buried in literature. It
also makes it difficult to translate results into applications, as
valuable results—relevant to the design of new drugs or new materials,
for example—are locked in publications. Moreover, these data are
loosely structured, encoded in manuscripts of various formats,
embedded in different content types, and are, in general, not machine
accessible. Thus, studies that automatically leverage this valuable
information are not tractable or even possible. This thesis studies
different approaches for liberating this information, exploring in
particular a hybrid human-computer solution that is able to
semi-automatically extract scientific facts from literature. This work
focuses on a particular field of molecular engineering as an example
scenario, although the approaches presented also apply to other
domains. The proposed approaches are implemented as χDB, a
framework to extract semi-structured properties from publications and
generate an associated database of curated molecular properties of
compounds. Such properties are frequently relied upon when designing
new materials, but are typically characterized only in published
literature rather than in databases. One example of such a property is
the Flory-Huggins interaction (χ) parameter, which measures the
interaction of polymer chains with solvent molecules as well as
polymer-polymer interactions. The goal of χDB is to liberate the
Flory-Huggins parameter from the literature and thus to enable
efficient access to Flory-Huggins parameter values for purposes of
comparison and material design. The first implementation uses web
crawling and crowdsourcing to extract χ values. Students from a
specially designed molecular engineering course curated publications
processed by the crawler and populated a database of χ values
using χDB, identifying hundreds of χ values not found in
previous survey articles. Lessons learned from student work with the
current system will guide future implementations.

Roselyne's advisor is Prof. Ian Foster

Login to the Computer Science Department website for details:
 https://www.cs.uchicago.edu/phd/ms_announcements#roselyne

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Margaret P. Jaffey            margaret at cs.uchicago.edu
Department of Computer Science
Student Support Rep (Ry 156)               (773) 702-6011
The University of Chicago      http://www.cs.uchicago.edu
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=