November 6, 2009
Ellen Riloff
Title: Toward the Automatic Acquisition of Semantic Taxonomies from Web Text
Semantic knowledge is enormously helpful for natural language processing, but difficult to obtain. Some manually created semantic resources exist, but they are far from complete and are often woefully inadequate for specialized domains. In this talk, we will describe bootstrapping algorithms that read Web texts to automatically acquire semantic taxonomic knowledge. The key to our approach is the use of a doubly-anchored hyponym pattern that can identify terms that participate in hypernym/hyponym (isa) relations (e.g., "dog" is a "mammal", and "mammal" is an "animal").
First, we will describe a bootstrapping algorithm that begins with one semantic class name (hypernym) (e.g., "mammal") and one seed hyponym term (e.g., "dog") and iteratively learns more hyponym terms that belong to the semantic class. This algorithm uses a graph structure to identify the most reliable hyponym terms. Second, we will describe a bootstrapping algorithm that uses the doubly-anchored hyponym pattern to learn new high-level conceptual terms (i.e., "vertebrate" or "predator") and iteratively use these new terms to fill out the concept space and generate taxonomic relations between words. Along the way, we will discuss some of the (immense!) challenges in formally evaluating the correctness of concepts and taxonomic relations.
(This is joint work with Eduard Hovy and Zornitsa Kozareva at the USC Information Sciences Institute.)
Posted by Admin at 04:30 PM
September 4, 2009
Adam Bargteil, Assistant Professor
Title: Physics-based Animation with Points
In this talk I will give an overview of physics-based animation and then discuss two recent papers on point-based animation of liquids and solids.
Posted by Admin at 04:13 PM | Comments (0)
September 11, 2009
Matt Might, Assistant Professor
Posted by Admin at 04:12 PM | Comments (0)
September 18, 2009
Valerio Pascucci, Associate Professor
CANCELLED
Posted by Admin at 04:12 PM | Comments (0)
September 25, 2009
Elaine Cohen, Professor & Rich Riesenfeld, Professor
1250 WEB
As part of the presentation ceremony, Richard Riesenfeld and Elaine Cohen, this year’s recipients of the international Pierre Bézier Award for Solid, Geometric and Physical Modeling and Applications, are giving a lecture on a topic of their choosing at 2009 SIAM/ACM Joint International Conference on Geometric and Physical Modeling, which will be held in San Francisco, 5-8 Oct 2009. Identical to the award lecture, this colloquium is being offered in advance of the San Francisco meeting so the interested local audience has the opportunity to hear it.
Pierre Bézier Award Lecture: Bézier, B-splines and Beyond
In an interleaved presentation drawing on many years of personal interactions with him, the recipients will begin by presenting some historical remarks concerning less known but important lessons learned from Pierre Bézier on how he achieved enormous impact. These certainly increased our appreciation and understanding for viewing mathematical contributions to CAD/CAM and modeling research in a larger context. This will help to highlight how his insight and work contributed inspiration for much of our subsequent work on B-spline methods applied to CAD/CAM, geometric modeling and computer graphics. By taking a long perspective in time on major advances in the field, we will try to assess where we are and indentify issues that call for redoubled efforts in the future. This analysis exposes that, in a somewhat ironic turn of events, we are now embarking on a recurring cycle of historical themes, a kind of wheel of reincarnation, albeit in a vastly more sophisticated and complex setting.
Posted by Admin at 04:11 PM | Comments (0)
October 2, 2009
Ross Whitaker, Professor
Title: Parameterizing high-dimensional data sets with kernel map manifolds
Many important data analysis problems come in in the form of a set of
data points each of which contains a large number of measurements,
which can be considered scattered data in a very high dimensional
space. Visualizing and analyzing such data is challenging, because
the dimensionality of the ambient space makes visualization and
statistical analysis quite difficult. However, often such data sets do
not fill the ambient space, but rather lie close to some lower
dimensional manifold. If the manifold is linear, then principal
component analysis and other linear models can extract the best
fitting models. However, the nonlinear case demands a more
sophisticated set of tools for learning the underlying structure of
high-dimensional data. This talk examines the problem of manifold
learning from a machine learning point of view and describes new tools
that make the connection between manifold learning and the statistical
generalization of PCA, called principal surfaces. We also present
results on examples of visualization and analysis of high-dimensional
data from graphics, perception, and medicine.
Posted by Admin at 04:11 PM | Comments (0)
October 23, 2009
Eric Eide and Rob Ricci
Title: Growing the Emulab Network Emulation Testbed
Emulab is a network emulation testbed that provides researchers with a wide
range of environments in which to develop, debug, and evaluate their systems.
Created by the Flux Research Group in 2000, it has continually expanded in size
and scope for the past decade. Today there are Emulab installations at more
than two dozen sites around the world. The Emulab site at Utah continues to be
a hub for systems research, with hundreds of users around the world running
thousands of experiments per year.
This talk will summarize recent and ongoing research projects toward continuing
Emulab's success as a premier facility for systems research. Emulab continues
to grow in scale toward support for experiments involving many thousands of
actual and virtual devices. It continues to improve with new support for
particular domains, such as cybersecurity and wireless networking. It
continues to grow as a basis for collaboration, and it is a control framework
for a prototype of the NSF's Global Environment for Network Innovation (GENI).
Finally, Emulab continues to expand in terms of the services it provides to
experimenters, with support for features such as live Internet conditions,
distributed checkpointing, and workflow management.
Posted by Admin at 03:09 PM