arrow_back
FW: Seminar on "Deep Learning" Friday 03/14 - FOR JEFFREY'S ATTENTION
1 message picture_as_pdf Source PDF
L
Lesley Groff Mar 14, 2008 7:44 PM
To
J. Epstein
Lesley Groff
Executive Assistant to Jeffrey
Epstein
From: egneurocog@aol.com [mailto:egneurocog@aol.com]
Sent: Thursday, March 13, 2008 7:16 PM
To: Lesley Groff; Lesley Groff
Subject: Fwd: Seminar on "Deep Learning" Friday 03/14 - FOR JEFFREY'S ATTENTION
I enjoyed the opportunity to meet and to clarify your vision of the project and your ideas how to make it work. Coincidentally, I received after our meeting an announcement of the talk that Yann is giving about his work tomorrow at Courant (see below). I plan to attend. Would you like to come?
Best wishes,
Nick
Elkhonon Goldberg, Ph.D., ABPP
Clinical Professor of Neurology
New York University School of Medicine
Diplomate
American Board of Professional Psychology
in Clinical Neuropsychology
315 West 57th Street
Suite 401
New York, NY 10019
USA
tel +1 212 541 6412
fax +1 212 765 7158
egneurocog@aol.com
elkhonon.goldberg@med.nyu.edu
www.elkhonongoldberg.com
> Courant Applied Math Seminar: http://www.cims.nyu.edu/ams/
> Friday 03/14 at 2:30 PM.
> Courant Institute, Warren Weaver Hall, Rm 1302.
> Host: David Cai <cai@cims.nyu.edu>
> Speaker: Yann LeCun, CIMS
>
> ________________________________________________________________
>
> Deep Learning
> Yann LeCun
> Computer Science Department,
> Courant Institute, NYU
>
> A long-term goal of Machine Learning research (ML) is to help solve
> highy complex "intelligent" tasks, such as visual perception, auditory
> perception, and language understanding. To reach that goal, the ML
> community must solve two problems: the Deep Learning Problem, and the
> Normalization Problem.
>
> The Deep Learning Problem is related to the difficulty of training
> "deep architectures" composed of many non-linear layers of trainable
> modules. There is considerable theoretical and practical evidence that
> complex tasks, such as invariant object recognition in vision, require
> "deep" architectures, composed of multiple non-linear layers of
> trainable functions. This is in contrasts with much of ML research of
> the last 10 years, which has primarily focused on "shallow" models
> that are essentially linear functions of the parameters to be learned.
>
> Deep learning amounts to optimizing a very highly non-convex function
> in a very high dimensional space. Several methods have recently been
> proposed to train (or pre-train) such deep architectures in an
> unsupervised fashion. Each layer of the deep architecture is composed
> of an encoder which computes a feature vector from the input, and a
> decoder which reconstructs the input from the features. A large
> number of such layers can be stacked and trained sequentially, thereby
> learning a deep hierarchy of features (or representations). Each
> layer is trained in an unsupervised fashion to minimize the
> reconstruction error under certain constraints on the features, such
> as sparsity. This class of learning methods is called "energy-based",
> because it amounts to shaping a high-dimensional energy landscape
> (a.k.a. an un-normalized log-likelihood function).
>
> A particular class of methods for deep energy-based unsupervised
> learning will be described that can learn sparse and overcomplete
> representations of data. When applied to natural image patches, the
> method produces filters similar to those found in the mammalian
> primary visual cortex. A hierarchical vision system that extracts
> high-level features suitable for computer vision applications can be
> produced by stacking multiple layers of this simple
> module. Applications to invariant object recognition in images, and
> visual navigation for mobile robots will be shown.
>
>
Elkhonon Goldberg, Ph.D., ABPP
Clinical Professor of Neurology
New York University School of Medicine
Diplomate
American Board of Professional Psychology
in Clinical Neuropsychology
315 West 57th Street
Suite 401
New York, NY 10019
USA
tel +1 212 541 6412
fax +1 212 765 7158
egneurocog@aol.com
elkhonon.goldberg@med.nyu.edu
www.elkhonongoldberg.com
> Courant Applied Math Seminar: http://www.cims.nyu.edu/ams/
> Friday 03/14 at 2:30 PM.
> Courant Institute, Warren Weaver Hall, Rm 1302.
> Host: David Cai <cai@cims.nyu.edu>
> Speaker: Yann LeCun, CIMS
>
> ________________________________________________________________
>
> Deep Learning
> Yann LeCun
> Computer Science Department,
> Courant Institute, NYU
>
> A long-term goal of Machine Learning research (ML) is to help solve
> highy complex "intelligent" tasks, such as visual perception, auditory
> perception, and language understanding. To reach that goal, the ML
> community must solve two problems: the Deep Learning Problem, and the
> Normalization Problem.
>
> The Deep Learning Problem is related to the difficulty of training
> "deep architectures" composed of many non-linear layers of trainable
> modules. There is considerable theoretical and practical evidence that
> complex tasks, such as invariant object recognition in vision, require
> "deep" architectures, composed of multiple non-linear layers of
> trainable functions. This is in contrasts with much of ML research of
> the last 10 years, which has primarily focused on "shallow" models
> that are essentially linear functions of the parameters to be learned.
>
> Deep learning amounts to optimizing a very highly non-convex function
> in a very high dimensional space. Several methods have recently been
> proposed to train (or pre-train) such deep architectures in an
> unsupervised fashion. Each layer of the deep architecture is composed
> of an encoder which computes a feature vector from the input, and a
> decoder which reconstructs the input from the features. A large
> number of such layers can be stacked and trained sequentially, thereby
> learning a deep hierarchy of features (or representations). Each
> layer is trained in an unsupervised fashion to minimize the
> reconstruction error under certain constraints on the features, such
> as sparsity. This class of learning methods is called "energy-based",
> because it amounts to shaping a high-dimensional energy landscape
> (a.k.a. an un-normalized log-likelihood function).
>
> A particular class of methods for deep energy-based unsupervised
> learning will be described that can learn sparse and overcomplete
> representations of data. When applied to natural image patches, the
> method produces filters similar to those found in the mammalian
> primary visual cortex. A hierarchical vision system that extracts
> high-level features suitable for computer vision applications can be
> produced by stacking multiple layers of this simple
> module. Applications to invariant object recognition in images, and
> visual navigation for mobile robots will be shown.
>
>
Supercharge your AIM. Get the AIM toolbar for your browser.