[Colloquium] Reminder: Guest Speakers @ TTI-C Today (4/10/06)

[Katherine Cumming]

**********TTI-C Guest Speakers Today***********
                                 April 10, 2006
        Presented by:  Toyota Technological Institute at Chicago
 
(1)
 
Speaker:  Xiaofeng Ren, University of California at Berkeley
Speaker's home page:  http://www.cs.berkeley.edu/~xren/
 
Date: Monday, April 10, 2006 
Location: TTI-C Conference Room, Part of Theory Seminar
Time: 10:00 am      
Title:   Probabilistic Models for Parsing Images
Abstract:
A grand challenge of computer vision is to understand and parse natural
images into boundaries, surfaces and objects.  To solve this problem we
would inevitably need to work with visual entities and cues of heterogeneous
nature, such as brightness and texture at low-level, contour and region
grouping at mid-level, and shape recognition at high-level.  Learning to
represent and incorporate these entities and cues, along with the complexity
of the visual world itself, calls for probabilistic models for image
parsing.  Many previous efforts in this line suffer from issues such as lack
of a compact representation, lack of scale invariance or lack of
comprehensive experimentation.  We describe a scale-invariant image
representation using piecewise linear approximations of contours and the
constrained Delaunay triangulation (CDT) for completing gradientless gaps.
On top of the CDT graph we develop conditional random fields (CRF) for
contour completion, figure/ground organization as well as object
segmentation.  Large datasets of human-annotated natural images are utilized
for both training and evaluation.  Our quantitative results are the first to
demonstrate the working of mid-level visual cues in general natural scenes.
The CDT/CRF framework enables efficient representation and inference of both
bottom-up and top-down information, hence applicable to various vision
problems.  We extend our work to joint object recognition and segmentation,
in particular finding people, in static images and video. 
 
(2)
 
Speaker:  Yves Lussier, The University of Chicago
Speaker's home page:  http://www.dbmi.columbia.edu/~yal7001/
 
Date: Monday, April 10, 2006 
Location: TTI-C Conference Room - Bioinformatics Seminar
Time:  2:00 pm
 
Title:   Collecting and Integrating Phenotypic Information from Multiple
Databases for Correlation with Genomics
Abstract:
The emerging field of systems biology integrates the study of the genome
with the study of phenotypes, which includes all of an organism's physical
characteristics, from proteins and molecular pathways to large-scale
physical characteristics such as height.  In contrast to the genomic
annotations that have been standardized extensively, the granularity and
compositionality of non-molecular phenotypic characters varies significantly
across species, scales of biology, and communities of researchers or
clinicians.  The goal of this research is to improve our understanding of
complex diseases in terms of their genetic and molecular causes and
hopefully to further individualize the prediction and the treatment of
diseases.  We have published original studies on computational methods that
automatically predict and organize related phenotypes between species and
across highly heterogeneous phenotypic datasets.  The methods are
"high-throughput" in the sense that large amounts of data are processed
automatically.  Previously this laborious and rate limiting task was
performed by expert curators.  At present, researchers have ample knowledge
of the genome and ample knowledge of the pathophysiology of disease, but
relatively little knowledge of the link between the two.  This issue is
referred to as the "phenotype gap." 
We hypothesize that, with these new methods, it will be possible to directly
relate and organize massive amounts of "phenotypic data" of important
diseases such as diabetes and hypertension to their genetic causes and thus
bridge the gap.
 
During this presentation, we will describe high throughput methods that
efficiently bridge the "phenotype gap", a required methodology for
large-scale comparative studies of phenotypes in systems biology and
medicine.  A recently published system, PhenoGO, will be presented to
illustrate the potential impact of the methods.  PhenoGO automatically
augments Gene Ontology (GO) Annotations with additional phenotypic context,
such as cell type, tissue type, organ, disease or morphology.  Since GO
annotations already provide a mapping between genes and their related sub
cellular functions and processes, the PhenoGO system enables to
computationally discriminate between different functions of a gene across
distinct cell types or tissue.  The PhenoGO system has a precision of 91%
and recall of 92% and maps to identifiers that are associated with different
phenotypic ontologies, including the UMLS, Cell Ontology, Mouse Anatomy,
NCBI taxonomy, and Mammalian Phenotype Ontology.
 
 
----------------------------------------------------------------------------
------
If you have questions, or would like to meet the speaker, please contact
Katherine at 773-834-1994 or kcumming at tti-c.org   
For information on future TTI-C talks and events, please go to the TTI-C
Events page:  http://www.tti-c.org/events.html.  TTI-C (1427 East 60th
Street, Chicago, IL  60637)
 
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20060410/d1f4c9cd/attachment.htm