- Monday, 23 November 2009, 11:00AM -- D.L. Pratt, 290 C
- Machine Learning Seminar
- Speaker: Amit Gruber, Dept. Computer Science, University of Toronto
Title: "Latent Topic Models for Hypertext"
Abstract: Latent topic models have been successfully applied as an unsupervised topic discovery technique in large document collections. With the proliferation of hypertext document collection such as the Internet, there has also been great interest in extending these approaches to hypertext (Cohn and Hofmann '01, Erosheva et al. '04). These approaches typically model links in an analogous fashion to how they model words - the document-link co-occurrence matrix is modeled in the same way that the document-word co-occurrence matrix is modeled in standard topic models.
We present a probabilistic generative model for hypertext document collections that explicitly models the generation of links. Specifically, links from a word w to a document d depend directly on how frequent the topic of w is in d, in addition to the in-degree of d. We show how to perform EM learning on this model efficiently. By not modeling links as analogous to words, we end up using far fewer free parameters and obtain better link prediction results.
Joint work with Michal Rosen-Zvi and Yair Weiss.
- Wednesday, 25 November 2009, 10:00AM -- Bahen Centre, Rm. 5256
- Computational Vision Seminar
- Speaker: Yanxi Liu, Penn State University
Title: "Symmetry"
Abstract: Symmetry is an essential mathematical concept, as well as a ubiquitous,
observable phenomenon in nature, science and
art. Either by evolution or by design, symmetry implies an efficiency coding
that makes it universally appealing,
especially so to computational science. Recognition and categorization of
symmetry and regularity is the first step
towards capturing the essential skeleton of a real world problem, while at
the same time minimizing computational
redundancy. However, symmetry group detection from real world data turns out
to be a challenging problem that has been
puzzling computer vision, computer graphics and psychology researchers for
decades. We explore a formal and
computational characterization of real world regularity using a hierarchical
model of symmetry groups as a theoretical
basis, embedded in a well-defined Bayesian framework. Such a formalization
simultaneously facilitates (1) a robust and
comprehensive algorithmic treatment of the whole regularity spectrum, from
regular (perfect symmetry), near-regular
(approximate symmetry), to various types of irregularities; (2) an effective
detection scheme for real world symmetries
and symmetry groups; and (3) a set of computational bases for measuring and
discriminating quantified regularities on
diverse data sets. Besides some theoretical background on crystallographic
groups in particular, I shall illustrate
various applications of computational symmetry in texture synthesis,
analysis, tracking, and manipulation; human gait
and activity recognition; symmetry-based dance analysis; grid-cell
clustering; automatic geo-tagging; and image
‘de-fencing’.
BIO:
Yanxi Liu received her B.S. degree in physics/electrical engineering in
Beijing and her Ph.D.
degree in computer science for group theory applications in robotics from
University of
Massachusetts (Amherst). Her postdoctoral training was at LIFIA/IMAG
(France). She also spent
one year at DIMACS (NSF center for Discrete Mathematics and Theoretical
Computer Science)
under an NSF research-education fellowship award. Dr. Liu was with the
research faculty in the
Robotics Institute (RI) of Carnegie Mellon University before she joined the
Computer Science
Engineering and Electrical Engineering departments of Penn State University
in Fall of 2006 as a
tenured faculty and the co-director of the lab for perception, action and
cognition (LPAC). Dr. Liu's
research interests span a wide range of applications including computer
vision, computer
graphics, robotics, human perception and computer aided diagnosis in
medicine, with two main
themes: computational symmetry/regularity and discriminative subspace
learning. Dr. Liu chaired
the First International Workshop on Computer Vision for Biomedical Image
Applications (CVBIA)
in conjunction with ICCV 2005. Dr. Liu served as an area chair or organizing
committee member
for CVPR08/MICCAI08/CVPR09, and has served as a multi-year chartered study
section member
for the US National Institute of Health (NIH). Dr. Liu is a senior member of
IEEE and the IEEE
Computer Society.
- Thursday, 26 November 2009, 2:00PM -- GB 220 - Galbraith Bldg
- Combinatorics Seminar
- Speaker: Pawel Pralat, Department of Mathematics, West Virginia University
Title: "Chasing robbers on random graphs"
Abstract: We study the vertex pursuit game of Cops and Robbers where
cops try to capture a robber on the vertices of the graph. The minimum
number of cops required to win on a given graph G is the cop number of
G. We present asymptotic results for the game of Cops and Robber played
on a random graph G(n,p) for a wide range of p=p(n). It has been shown
that the cop number as a function of an average degree forms an
intriguing zigzag shape.
- Friday, 27 November 2009, 11:00AM -- GB 244 - Galbraith Bldg
- Theoretical Computer Science Seminar
- Speaker: Irénée Briquel, Laboratoire de l’Informatique du Parallélisme - ENS, Lyon and Fields Institure
Title: "Lower bounds on the tree-width of boolean formulas."
Abstract: To a boolean formula can be associated the clause graph, where the
vertices are the variables, and where two variables are linked in the
graph when they belong to the same clause.
In a previous work, Pascal Koiran and Klaus Meer studied the link
between the complexity of the formula and the tree-width of the clause
graph - for short, the tree-width of the formula. They found an algorithm
to compute efficiently polynomials associated with boolean formulas of
bounded tree-width.
To estimate the limits of this method, it is interesting to look for
lower bounds on the tree-width of boolean formulas, establishing that the
previously mentioned algorithm is not efficient for the associated
polynomials.
To find lower bounds on the tree-width, we show a link between this
notion and the communication complexity of the boolean formula. We show
that lower bounds on the communication complexity can transpose to the
tree-width, and explore the possibilities of this method.
This is based on joint work with Pascal Koiran and Klaus Meer.
- Tuesday, 1 December 2009, 11:00AM -- Bahen Centre, Rm. 1180
- Distinguished Lecture Series Lecture
- Speaker: Professor Fran Allen, IBM T.J. Watson Research Center
Title: "High Performance Computers and Compilers: A Personal Perspective"
Remarks: Fran Allen, IBM Fellow Emerita
IBM T. J. Watson Research Center
1101 Kitchawan Rd.
Yorktown Heights, NY 10598
Abstract: The talk will describe a related sequence of projects including some early, very bold projects that profoundly influenced high performance computing even as some of them failed. The talk includes a personal perspective of what worked and what didn’t, the historical threads of some ideas and the lessons learned. The talk concludes by identifying some current compiler challenges and the need for a new focus on new compilers.
- Friday, 4 December 2009, 11:00AM -- D.L.Pratt Building, Rm. 266
- Computational Vision Seminar
- Speaker: Sam Hasinoff, MIT CSAIL
Title: "Fragmented Lenses and High ISO for Efficient Photography"
Abstract: I'll describe two projects that address basic technical challenges in
photography: (1) minimizing defocus blur, and (2) capturing high
dynamic range. In both cases we characterize fundamental limits, and
propose new methods which improve efficiency over the
state-of-the-art.
First, I'll describe our new lens design, the "lattice-focal" lens,
that can capture in-focus images over a greater range of depths than
previous approaches. The design follows from our analysis of lens
defocus over the 4D space of light rays. As we show, the only usable
energy lies on a 3D subset of this space in the Fourier domain. We
establish an upper bound on performance (ie. over any possible lens
design), and show that the lattice-focal lens is closer to this bound
than any previous design.
Second, I'll show how existing cameras can be used more efficiently,
to capture high dynamic range scenes. For a given scene and camera,
our analysis lets us compute the optimal sequence of photos to
capture, maximizing worst-case SNR. This provides significant gains
over standard exposure bracketing, typically 10 dB better or 3 times
faster, when capture time is limited. As I'll explain, most our gains
come from using high (but varying) ISO settings -- counterintuitively,
"turning up the amplifier" can help reduce noise.
Bio:
Sam Hasinoff received the BSc degree in computer science from the
University of British Columbia in 2000, and the MSc and PhD degrees in
computer science from the University of Toronto in 2002 and 2008,
respectively. He is currently an NSERC Postdoctoral Fellow at the
Massachusetts Institute of Technology. In 2006, he received an
honorable mention for the Longuet-Higgins Best Paper Award at the
European Conference on Computer Vision. He is the recipient of the
Alain Fournier Award for the top Canadian dissertation in computer
graphics in 2008.
http://www.csail.mit.edu/~hasinoff/
- Tuesday, 9 February 2010, 11:00AM -- Bahen Centre, Rm. 1180
- Distinguished Lecture Series Lecture
- Speaker: Dr. Joe Marks, Vice President, Disney Research
Title: "Interactive Media Research at The Walt Disney Company"
Remarks: Speaker bio: Joe Marks grew up in Dublin, Ireland. Being more perseverant than imaginative, he earned three degrees from Harvard University. His areas of interest include computer graphics, human-computer interaction, and artificial intelligence. He has worked previously at Bolt Beranek and Newman and at Digital's Cambridge Research Laboratory. Prior to joining The Walt Disney Company he was the Research Director at Mitsubishi Electric Research Labs in Cambridge, MA, from 2000-2006.
Abstract: The research challenges in interactive media faced by the different business units of The Walt Disney Company fall in four broad categories: motion pictures, park attractions, games & sports, and media networks. Enumerating these challenges provides a key industry perspective on promising directions for future research in digital arts, experiential media systems, creative environments, and emerging media technologies.
- Tuesday, 30 March 2010, 11:00AM -- Bahen Centre, Rm. 1180
- Distinguished Lecture Series Lecture
- Speaker: Professor Maja Mataric, Computer Science and Neuroscience, University of Southern California
Title: "To be announced"