REMINDER: UdeM-McGill-MITACS machine learning seminar, PAA 1409 (UdeM), Fri Oct 19, 11:30am
This week's seminar (see http://www.iro.umontreal.ca/article.php3? id_article=107&lang=en):
Learning the 2-D Topology of Images
by Pascal Lamblin, Département d’Informatique et Recherche Opérationnelle Université de Montréal
Location: Pavillon André-Aisenstadt (UdeM), room 1409 Time: October 19th 2007, 11h30
We study the following question: is the two-dimensional structure of images a very strong prior or is it something that can be learned with a few examples of natural images? If someone gave us a learning task involving images for which the two-dimensional topology of pixels was not known, could we discover it automatically and exploit it? For example suppose that the pixels had been permuted in a fixed but unknown way, could we recover the relative two-dimensional location of pixels on images?
The surprising result presented here is that not only the answer is yes but that about as few as a thousand images are enough to approximately recover the relative locations of about a thousand pixels. This is achieved using a manifold learning algorithm applied to pixels associated with a measure of distributional similarity between pixel intensities. We compare different topology-extraction approaches and show how having the two-dimensional topology can be exploited.
Afficher les réponses par date
Bonjour,
J'ai pensé que cela pourrait sûrement intéresser certaines personnes du lab. Désolé pour l'annonce tardive, la conférence a lieu demain à 15h30 à McGill.
Bonne soirée,
Sébastien
From: CRM activites@CRM.UMontreal.CA To: lofficiel lofficiel@CRM.UMontreal.CA Subject: [CRM-L'officiel] COLLOQUE CRM-ISM-GERAD DE STATISTIQUE: R. Nugent (Le vendredi 26 octobre 2007)
RAPPEL - REMINDER - RAPPEL - REMINDER - RAPPEL
****************************************************************** COLLOQUE CRM-ISM-GERAD DE STATISTIQUE ******************************************************************
CONFERENCIER(S) / SPEAKER(S) : Rebecca Nugent (Carnegie Mellon University)
TITRE / TITLE : "Visualizing Clusters With a Density-Based Similarity Measure"
LIEU / PLACE : McGill, Burnside Hall, 805 Sherbrooke O., BH 1B45
DATE : Le vendredi 26 octobre 2007 / Friday, October 26, 2007
HEURE / TIME : 15 h 30 / 3:30 p.m.
RESUME / ABSTRACT : The goal of clustering is to identify distinct groups in a dataset and assign a group label to each observation. To cast clustering as a statistical problem, we regard the data as a sample from an unknown density p(x). To generate clusters, we estimate the properties of p(x) either with parametric (model-based) or nonparametric methods. In contrast, the algorithmic approach to clustering (linkage methods, spectral clustering) applies an algorithm, often based on a distance measure, to data in m-dimensional space. Many commonly used clustering methods employ functions of Euclidean distance between observations to determine groupings. Spherical groups are easily identified, curvilinear groups less so. We first motivate the use of a density-based similarity measure and briefly introduce generalized single linkage, a graph-based clustering approach. We describe a refinement algorithm used to bound the measure and then explore the performance of this measure in clustering and visualization methods.
****************************************************************** http://www.crm.umontreal.ca/cgi/Stats.csh
lisa_seminaires@iro.umontreal.ca
-
Hugo Larochelle -
Sébastien Gambs