Skip navigation.
Semantic Software Lab
Concordia University
Montréal, Canada


CSE Seminar: Developments in general purpose GPU computing

2012-11-05 15:00
2012-11-05 16:00

Speaker: Dr. Ming Ouyang, Computer Engineering & Computer Science Department, University of Louisville
Date: Monday, November 5, 2012, 15:00
Room: EV3.309, Concordia SGW

1. Abstract

Graphics processing units (GPUs) on commodity video cards were originally designed towards the needs of the 3-D gaming industry for high performance, real-time graphics. They have become powerful co-processors to the CPUs. The top of the line Nvidia GPUs for computation have 512 cores in one chip. Scientists and engineers from many disciplines are exploring various ways to use this massive amount of parallel computation. This presentation gives an introduction of GPU hardware and programming, and a survey of some applications.

CSE Seminar Talk, Dr. Stephann Makri, "Coming Across Information Serendipitously: An Empirical Study", Concordia University, Montréal

2012-05-01 11:00
2012-05-01 12:00

1. Date & Place

May 1st, 2012, 11am-12noon, Concordia University, Montréal, SGW Campus, EV3.309

2. Abstract

We wanted to gain a detailed empirical understanding of how
researchers come across information serendipitously, grounded in
real-world examples. To gain this understanding, we asked 28
researchers from a broad cross-section of disciplines to discuss in
detail memorable examples of coming across information serendipitously
from their research or everyday life. We found that although the
examples provided were varied, they shared common elements
(specifically, they involved a mix of unexpectedness and insight and
led to a valuable, unanticipated outcome). These elements form the
core of 1) a descriptive model of serendipity and 2) a framework for
subjectively classifying whether or not a particular experience might
be considered serendipitous and, if so, how serendipitous. In this
talk, we discuss this model and framework and the implications of our
findings on the design of interactive systems.

CSE Seminar Talk, Dr. Rudzicz, "First, we shape our tools: How to build a better speech recognizer", Concordia University, Montréal

2011-12-02 14:00
2011-12-02 15:00

1. Abstract

In this talk I briefly survey some of my previous research and then even more briefly extrapolate as to future extensions of this work.

I will talk about improving Automatic Speech Recognition (ASR) for speakers with speech disabilities by incorporating knowledge of their speech production. This involves the acquisition of the TORGO database of disabled articulation which demonstrates several consistent behaviours among speakers, including predictable pronunciation errors. Articulatory data are then used to train augmented ASR systems that model the statistical relationships between the vocal tract and its acoustic effluence. I show that dynamic Bayesian networks augmented with instantaneous articulatory variables outperform even discriminative alternatives. This leads to work that incorporates a more rigid theory of speech production, i.e., task-dynamics, that models the high-level and long-term aspects of speech production. For this task, I devised an algorithm for estimating articulatory positions given only acoustics that significantly outperforms the former state-of-the-art.

Finally, I present ongoing work into the transformation of disabled speech signals in order to make them more intelligible to human listener and I conclude with some thoughts as to possible paths we may now take.

Syndicate content