[Colloquium] Guest Speakers @ TTI-C Next Week (3/13/06-3/17/06)

[Katherine Cumming]

**********TTI-C Guest Speakers (4) Next Week***********
                      March 13 - March 17, 2006
        Presented by:  Toyota Technological Institute at Chicago
 
(1)
 
Speaker: Dan L. Nicolae, Department of Statistics, University of Chicago
Speaker's home page:  <http://www.stat.uchicago.edu/faculty/nicolae.html>
http://www.stat.uchicago.edu/faculty/nicolae.html
 
Date: Monday, March 13, 2006 
Location: TTI-C Conference Room, Part of Bioinformatics Seminar
Time:  2:00 pm 
         
Title:  Whole-genome Association SNP Sets
Abstract:
Large-scale genotyping platforms are used in genetic association 
Studies that aim to identify the DNA variation affecting risk to disease, or
modifying the distribution of phenotypic variables such as quantitative
traits and response to drug.  The Affymetrix genotyping sets, GeneChip
Mapping 100K and 500K Arrays, are among the most used platforms in
current studies.  I will talk about issues in study design and data
analyses and I will describe genotype-calling algorithms for these
arrays.
 
(2)
 
Speaker:  Brona Brejova, University of Waterloo
Speaker's home page:  <http://www.bioinformatics.uwaterloo.ca/~bbrejova/>
http://www.bioinformatics.uwaterloo.ca/~bbrejova/
 
Date: Tuesday, March 14, 2006 
Location: TTI-C Conference Room
Time:  10:00 am
 
Title:   Knowing More about Alignments Helps in Finding Them
Abstract:
The local alignment problem is the task of finding pairs of similar
substrings in a large DNA or protein sequence database.  It is one of the
basic steps in analyzing a new sequence.  Alignments of similar sequences
highlight areas preserved by evolution and help to determine structure and
biological function of sequences.  Increasing amount of available genomic
sequence data creates the need for further improvements in efficiency and
accuracy of local alignment tools.

We explore the technique of spaced seeds (developed by Ma, Tromp, and Li in
2002) for increasing the accuracy of a commonly used heuristic local
alignment algorithm by tuning a certain parameter called spaced seed.  We
argue that the spaced seed should be optimized with respect to a realistic
probabilistic model of biologically meaningful alignments.  The seeds that
we discover significantly improve the accuracy and running time of local
alignment for protein coding regions in DNA.  We also generalize the notion
of spaced seed to encompass several approaches suggested by other
researchers.  This unifying framework allows the extension of our ideas to
protein sequences.

This is joint work with Dan Brown and Tomas Vinar. 
 
(3)
 
Speaker:  Silvio Micali, MIT CSAIL
Speaker's home page:
<http://www.csail.mit.edu/biographies/PI/bioprint.php?PeopleID=36>
http://www.csail.mit.edu/biographies/PI/bioprint.php?PeopleID=36
 
 
Date: Thursday, March 16, 2006 
Location:  Ryerson 251, Part of TTI-C Distinguished Lecturer Series
Time:  3:00 pm
 
Title:   Rational Secure Computation and Ideal Mechanism Design
 
Abstract: 
We prove a general result bridging the fields of Secure Protocols and Game
Theory.

In game-theoretic terms, we show that ANY mediated game with incomplete
information can be perfectly simulated by the players alone, by means of an
extensive-form game in which the trusted mediator is replaced by a ballot
box---the venerable device used throughout the world to privately and
correctly compute the tally of secret votes. 

In cryptographic terms, we show that, in ANY joint computation, security can
be achieved based solely on the players' RATIONALITY, rather than on the
HONESTY of some of them. 

Our result has broad implications for Mechanism Design; in particular, it
enables One to design mechanisms in a MODULAR and COMPETITIVE fashion. 

Joint work with Sergei Izmalkov and Matt Lepinski 
 
(4)
 
Speaker: Nathan Srebro, University of Toronto
Speaker's home page:  <http://people.csail.mit.edu/nati/>
http://people.csail.mit.edu/nati/
 
 
Date: Friday, March 17, 2006 
Location: TTI-C Conference Room
Time:  10:00am
Title:  Maximum Margin Matrix Factorization
Abstract:
Factor, or linear component (PCA), models are often natural in the analysis
of many kinds of tabulated data (e.g. collections of documents or images,
gene expression measurements and user preferences). The premise of such
models is that important aspects of the data can be captured by a small
number dimensions ("components", "factors" or "topics").

I will present a novel approach that allows an unbounded (infinite) number
of factors. This is achieved by limiting the norm of the factorization
instead of its dimensionality. The approach is inspired by, and has strong
connections to, large-margin linear discrimination.

I will show how such a max-margin matrix factorization can be learned by
solving a (very large, but efficiently solvable) semi-definite program. I
will also present generalization error bounds for learning with such
factorization, and discuss the relationship between what can be learned with
max-margin and low-dimensional factorizations.

Joint work with Alexandre d'Aspremont, Tommi Jaakkola and Jason Rennie and
Adi Shraibman. 
 
 
Abstract:
----------------------------------------------------------------------------
------
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/20060310/dbaf9dc1/attachment.htm