October 2008 - Lisa_seminaires - mailman.iro.umontreal.ca

FW: CERNEC Conférence - Vendredi 17 Oct. (Dr. Marc Schoenwiesner)
by Douglas Eck 14 Oct '08

14 Oct '08

Here's a really interesting talk. Unfortunately it conflicts with Geoff Hinton's talk. . . . but I wanted to announce it anyway. -Doug ---------- Forwarded message ---------- From: Schoenwiesner Marc <marc.schoenwiesner(a)umontreal.ca> Date: 2008/10/14 Subject: [BRAMS-local] FW: CERNEC Conférence - Vendredi 17 Oct. (Dr. Marc Schoenwiesner) To: brams-local(a)iro.umontreal.ca Bonjour, Vous êtes cordialement invités à la prochaine conférence du Centre de Recherche en Neuropsychologie et Cognition (CERNEC). Hi, You are cordially invited to the next Centre de Recherche en Neuropsychologie et Cognition (CERNEC) conference. Conférencier / Speaker : Dr. Marc Schoenwiesner 'The human hearing system - what we know, what we want to know, and how we can find out' Date : Vendredi 17 octobre 2008, 15h00 Lieu : Local D-427 du pavillon Marie-Victorin de l'Université de Montréal, 90 avenue Vincent d'Indy (à la sortie du métro Édouard-Montpetit)

1 0

Oct 16th, 15:00 - UdeM-McGill-MITACS seminar by Geoffrey Hinton
by Pascal Lamblin 10 Oct '08

10 Oct '08

The second MITACS seminar of the fall will be on Thursday, October 16th, 15:00 at UdeM. Title: Visualizing high-dimensional data using t-SNE Speaker: Geoffrey Hinton Date: Thursday Oct 16th, 2008, 15:00 Place: Room 3195 (pav. André-Aisenstadt) Abstract: Over the last decade, many new methods have been developed for visualizing high-dimensional data by giving each data-point a location in a two-dimensional map. The goal is to represent the separations of pairs of data-points by the separations of their corresponding map-points, with an emphasis on representing the small separations accurately. I will describe a new method, called t-SNE, that is based on two ideas. The first idea is to convert the set of pairwise distances between data-points into a set of probabilities of selecting pairs of data-points. The selection probability of a pair of points is proportional to a Gaussian function of their separation. If the distances between map-points are converted into pairwise probabilities in the same way, any given arrangement of map-points can be evaluated by measuring the divergence between the probability distributions obtained from the data-points and the map-points. A good arrangement of map-points is then found by performing gradient descent in this divergence. Unfortunately, if the probabilities of pairs of map-points are computed using a Gaussian function of their separation, the difference between the distributions of pairwise distances in high-dimensional and low-dimensional spaces causes the map-points to be crowded together in the center of the map. This problem can be largely overcome by using a heavy-tailed t-distribution when computing the selection probabilities of pairs of map-points. This leads to maps that look much better than those produced by other recent methods. In particular, t-SNE is very good at preserving clusters in the data at many different scales simultaneously. The talk describes joint work with Laurens van der Maaten that will appear in the Journal of Machine Learning Research. -- Pascal

1 1

Oct 17th, 14:30 talk by Hinton at McGill - The next generation of neural networks
by Pascal Lamblin 07 Oct '08

07 Oct '08

Geoff Hinton is giving a talk next week (friday, oct. 17th) at McGill, organized by CRM Applied Math Lab and McGill CSE, <http://www.dms.umontreal.ca/~mathapp/> Location: McGill, Burnside Hall 1205 Time: 14:30 Speaker: Geoffrey Hinton <http://www.cs.toronto.edu/~hinton/> Title: THE NEXT GENERATION OF NEURAL NETWORKS Abstract: In the 1980's, new learning algorithms for neural networks promised to solve difficult classification tasks, like speech or object recognition, by learning many layers of non-linear features. The results were disappointing for two reasons: There was never enough labeled data to learn millions of complicated features and the learning was much too slow in deep neural networks with many layers of features. These problems can now be overcome by learning one layer of features at a time and by changing the goal of learning. Instead of trying to predict the labels, the learning algorithm tries to create a generative model that produces data which looks just like the unlabeled training data. After learning many layers of features in this way, a relatively small amount of labeled data can then be used to fine-tune the features to give better discrimination. These new neural networks outperform other machine learning methods when labeled data is scarce but unlabeled data is plentiful. I will describe an application to recognizing 3-D shapes by Vinod Nair and an application to generating human motion by Graham Taylor. -- Pascal

1 0

Oct 10th, 14:00 - UdeM-McGill-MITACS seminar by Ronan Collobert
by Pascal Lamblin 07 Oct '08

07 Oct '08

Ladies and gentlemen, Here comes the first MITACS seminar, by Ronan Collobert, this friday, 2 pm at UdeM. Ronan Collobert Friday Oct 10th, 2008, 14:00 Room 3195 (pav. André-Aisenstadt) Large Scale Learning for Natural Language Processing We describe a neural network architecture that given a sentence, outputs a host of language processing predictions: part-of-speech tags, chunks, named entity tags, semantic roles, semantically similar words and the likelihood that the sentence makes sense (grammatically and semantically) using a language model. The entire network is trained jointly on all these tasks using weight-sharing, an instance of multitask learning. All the tasks use labeled data except the language model which is learnt from unlabeled text and represents a novel way of performing semi-supervised learning for the shared tasks. We show how both multitask learning and semi-supervised learning improve the generalization of the shared tasks, resulting in a learnt model with state-of-the-art performance. -- Pascal

1 0