deep learning & feature learning methods for vision rob fergus (nyu) kai yu (baidu) marc’aurelio...

67

Deep Learning & Feature Learning Methods for Vision Rob Fergus (NYU) Kai Yu (Baidu) Marc’Aurelio Ranzato (Google) Honglak Lee (Michigan) Ruslan Salakhutdinov (U. Toronto) Graham Taylor (University of Guelph) CVPR 2012 Tutorial: 9am-5:30pm

Upload: stewart-douglas

Post on 18-Dec-2015

220 views

Category:

Documents

4 download

Report

Download

Tags:

Embed Size (px):

TRANSCRIPT

Slide 1
Deep Learning & Feature Learning Methods for Vision Rob Fergus (NYU) Kai Yu (Baidu) MarcAurelio Ranzato (Google) Honglak Lee (Michigan) Ruslan Salakhutdinov (U. Toronto) Graham Taylor (University of Guelph) CVPR 2012 Tutorial: 9am-5:30pm
Slide 2
Tutorial Overview 9.00am:IntroductionRob Fergus (NYU) 10.00am:Coffee Break 10.30am:Sparse CodingKai Yu (Baidu) 11.30am:Neural NetworksMarcAurelio Ranzato (Google) 12.30pm:Lunch 1.30pm:Restricted Boltzmann Honglak Lee (Michigan) Machines 2.30pm:Deep BoltzmannRuslan Salakhutdinov (Toronto) Machines 3.00pm:Coffee Break 3.30pm:Transfer Learning Ruslan Salakhutdinov (Toronto) 4.00pm:Motion & Video Graham Taylor (Guelph) 5.00pm:Summary / Q & AAll 5.30pm:End
Slide 3
Overview Learning Feature Hierarchies for Vision Mainly for recognition Many possible titles: Deep Learning Feature Learning Unsupervised Feature Learning This talk: Basic concepts Links to existing vision approaches
Slide 4
Existing Recognition Approach Hand- designed Feature Extraction Trainable Classifier Image/Video Pixels Features are not learned Trainable classifier is often generic (e.g. SVM) Object Class Slide: Y.LeCun
Slide 5
Motivation Features are key to recent progress in recognition Multitude of hand-designed features currently in use SIFT, HOG, LBP, MSER, Color-SIFT. Where next? Better classifiers? Or keep building more features? Felzenszwalb, Girshick, McAllester and Ramanan, PAMI 2007 Yan & Huang (Winner of PASCAL 2010 classification competition)
Slide 6
What Limits Current Performance? Ablation studies on Deformable Parts Model Felzenszwalb, Girshick, McAllester, Ramanan, PAMI10 Replace each part with humans (Amazon Turk) : Also removal of part deformations has small (

Analyzing Uncertainty in Neural Machine Translation · Analyzing Uncertainty in Neural Machine Translation Myle Ott 1Michael Auli David Grangier Marc’Aurelio Ranzato1 Abstract Machine

Discrete Restricted Boltzmann Machinesjmlr.csail.mit.edu/papers/volume16/montufar15a/montufar15a.pdfDiscrete Restricted Boltzmann Machines 2008; Ranzato et al., 2010). The techniques

Deep Visual Analogy-Making - University of Michiganweb.eecs.umich.edu/~honglak/nips2015-analogy.pdf · Deep Visual Analogy-Making Scott Reed Yi Zhang Yuting Zhang Honglak Lee University

Lecture 06 marco aurelio ranzato - deep learning

Gradient Episodic Memory for Continual Learning - arXiv · PDF fileGradient Episodic Memory for Continual Learning David Lopez-Paz and Marc’Aurelio Ranzato Facebook Artiﬁcial Intelligence

Unsupervised Learning of Feature Hierarchiesranzato/publications/ranzato-phd...Unsupervised Learning of Feature Hierarchies by Marc’Aurelio Ranzato A dissertation submitted in partial

DEFTCON 2014 – Luigi Ranzato - Windows Registry Artifacts: "Most Recently Used" Shellbags

A shallow look at Deep Learning Computer Vision James Hays Many slides from CVPR 2014 Deep Learning Tutorial (Honglak Lee and Marc’Aurelio especially)

Deep Learning Part I · Slides Credit: Deep Learning Tutorial by Marc’Aurelio Ranzato • Each convolution kernel is a local pattern detector • Use many of them in a convolutional

PANDA: Pose Aligned Networks for Deep Attribute Modeling Ning Zhang1;2, Manohar Paluri1, Marc’Aurelio Ranzato1, Trevor Darrell2, Lubomir Bourdev1 1: Facebook

Automatic Single-Image 3d Reconstructions of Indoor ...web.eecs.umich.edu/~honglak/isrr05_3dReconManScenes.pdf · Automatic Single-Image 3d Reconstructions of Indoor Manhattan World

Learning Factored Representations in a Deep …ranzato/publications/eigen_iclr14.pdfLearning Factored Representations in a Deep Mixture of Experts David Eigen 1;2 Marc’Aurelio Ranzato

Weakly Supervised Semantic Segmentation using Web-Crawled ...web.eecs.umich.edu/~honglak/cvpr17-weaksupSegmWebVideos.pdf · Weakly Supervised Semantic Segmentation using Web-Crawled

Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representations Honglak Lee, Roger Grosse, Rajesh Ranganath, and

Generating more realistic images using gated MRF's...Generating more realistic images using gated MRF’s Marc’Aurelio Ranzato Volodymyr Mnih Geoffrey E. Hinton Department of Computer

UNSUPERVISED FEATURE LEARNING VIA SPARSE ...web.eecs.umich.edu/~honglak/thesis_final.pdfUNSUPERVISED FEATURE LEARNING VIA SPARSE HIERARCHICAL REPRESENTATIONS A DISSERTATION SUBMITTED

Anton Bakhtin Sam Gross Myle Ott Yuntian Deng Marc’Aurelio ... · Anton Bakhtin Sam Gross Myle Ott Yuntian DengF Marc’Aurelio Ranzato Arthur Szlam Facebook AI Research FHarvard

Factored 3-Way Restricted Boltzmann Machines For ...hinton/absps/ranzato_aistats2010.pdfFactored 3-Way Restricted Boltzmann Machines For Modeling Natural Images Marc’Aurelio Ranzato

An Introduction to Deep Learningranzato/files/ranzato_deeplearn17_lec2_nlp.pdf · An Introduction to Deep Learning Marc’Aurelio Ranzato Facebook AI Research [email protected] DeepLearn

Large-Scale Visual Recognition With Deep …ranzato/publications/ranzato_cvpr13.pdfLarge-Scale Visual Recognition With Deep Learning Sunday 23 June 2013 Marc'Aurelio Ranzato [email protected]

Understanding and Improving Convolutional Neural Networks ...web.eecs.umich.edu/~honglak/icml2016-crelu-full.pdfUnderstanding and Improving Convolutional Neural Networks via Concatenated

arXiv:1812.00420v2 [cs.LG] 9 Jan 2019 · 2019-01-10 · Published as a conference paper at ICLR 2019 EFFICIENT LIFELONG LEARNING WITH A-GEM Arslan Chaudhry1, Marc’Aurelio Ranzato

DistBelief: Joint work with: Kai Chen, Greg Corrado, Jeff Dean, Matthieu Devin, Rajat Monga, Andrew Ng, Marc’Aurelio Ranzato, Paul Tucker, Ke Yang Thanks:

Structured Recurrent Temporal Restricted Boltzmann Machinesweb.eecs.umich.edu/~honglak/icml2014_structured-rtrbm.pdf · 2014-05-30 · Structured Recurrent Temporal Restricted Boltzmann

UNSUPERVISED FEATURE LEARNING VIA …honglak/thesis_final.pdfUNSUPERVISED FEATURE LEARNING VIA SPARSE HIERARCHICAL REPRESENTATIONS ... and having a poor representation can ... 4.2.2

Generating more realistic images using gated MRF’sfritz/absps/mcimage.pdf · Generating more realistic images using gated MRF’s Marc’Aurelio Ranzato Volodymyr Mnih Geoffrey

Unsupervised learning of hierarchical representations with ...web.eecs.umich.edu/~honglak/cacm2011-researchHighlights-convDBN.pdf · Hierarchical Representations with Convolutional

Large Memory Layers with Product Keys · 2019-12-17 · Large Memory Layers with Product Keys Guillaume Lampley, Alexandre Sablayrolles , Marc’Aurelio Ranzato , Ludovic Denoyery,

Unsupervised Deep LearningUnsupervised Deep Learning Tutorial - Part 2 Alex Graves Marc’Aurelio Ranzato NeurIPS, 3 December 2018 [email protected] [email protected] Overview • Practical

Modeling Pixel Means and Covariances Using Factorized ...ranzato/publications/ranzato_cvpr2010.pdf · Using Factorized Third-Order Boltzmann Machines Marc’Aurelio Ranzato Geoffrey

Learning to Disentangle Factors of Variation with Manifold …web.eecs.umich.edu/~honglak/icml2014_disentangling_final.pdf · Boltzmann machine that incorporates multiplica-tive interactions

Image Classification with Deep Learningranzato/files/ranzato_CNN... · Image Classification with Deep Learning Marc'Aurelio Ranzato Facebook A.I. Research ... Deep (sparse/denoising)

Image Classification with Deep Learningranzato/files/ranzato_CNN_stanford2015.pdfImage Classification with Deep Learning ... Traditional Pattern Recognition VISION SPEECH NLP Ranzato

Deep Visual Analogy-Making Scott Reed Yi Zhang Yuting Zhang Honglak Lee University of Michigan, Ann Arbor

How to generate realistic images using gated MRF’sranzato/publications/ranzato_nips...How to generate realistic images using gated MRF’s Marc’Aurelio Ranzato Volodymyr Mnih Geoffrey