deep networks for computer vision at google
play

Deep Networks for Computer Vision at Google Chuck Rosenberg - PowerPoint PPT Presentation

Dec 16, 2023 •230 likes •506 views

Deep Networks for Computer Vision at Google Chuck Rosenberg ImageNet ILSVRC Workshop September 12, 2014 Quick Intro Private Photo Search and Public Image Search Teams Google Photos Our work: Pixels Knowledge Search by Image

  1. Deep Networks for Computer Vision at Google Chuck Rosenberg ImageNet ILSVRC Workshop September 12, 2014

  2. Quick Intro Private Photo Search and Public Image Search Teams Google Photos Our work: Pixels → Knowledge

  3. Search by Image Applications Google Image Search

  4. Applications - Photo Search

  5. Applications - Auto Curation

  6. Google Photos - Auto Awesome

  7. More Image Understanding at Google YouTube Google Shopping Much Advertising more... StreetView / Maps Self-Driving Cars Robotics

  8. Understanding is about extracting Knowledge

  9. Image Understanding: Pixels → Entities Single-word entities go way beyond simple objects! My photos of … objects: “dog” fine-grained objects: “Husky” scenes: “beach”, “sunset” actions: “kitesurfing” ,“kiss” emotions: “happiness”, “laughter” events: “birthday”, “basketball game” abstract concepts: “love”, “zen”

  10. The Deep and now Deeper Hammer Target output Pixels Deep Neural Network Deep learning infrastructure by the Google Brain team “ImageNet Classification with Deep Convolutional Neural Networks”, Krizhevsky, Sutskever, Hinton, NIPS 2012

  11. Personal Photos - Example Annotations Crowd Hummingbird Play Christmas tree Cheering Macro photography Meal Red People Reflection Cake Christmas decoration Stadium Red Child Christmas

  12. More Example Network Annotations

  13. More Example Network Annotations

  14. Google Network Stats ● Training Data ○ ImageNet 1K ~1M images ○ X-Net 100’s of millions of images ● Label Set ○ Image 1K ○ X-Net ~10’s of thousands of labels ● Ground Truth Issues ○ Incomplete Training Data ○ Noisy Training Data

  15. Challenge: Incomplete label ground-truth Problem increasingly serious as we add more types of entities and fine- grained categories: “Airedale Terrier” but not “Terrier” “Dog” “Animal” or “Pet” “Cute” or “Curb” “Grass” “Street” ...

  16. Challenge: Noisy data “Tortoise Shell “Random noise” “Tortoise” Sunglasses”

  17. Image Understanding: Localization sky Mountain human running grass dog road object detection scene parsing pose estimation

  18. Sample detections ImageNet Pascal VOC

  19. Training Embeddings Using Triplets E Triplets M B E Triplet D Deep Neural Net L2 Loss D ... I N G ● Training data consists of triplets: an anchor image, positive image, and negative image. Negative ● Loss function: Anchor [1] Positive [1] “Learning Fine-grained Image Similarity with Deep Ranking”, Wang, Song, Leung, Rosenberg, Wang, Philbin, Chen, Wu, CVPR 2014 Google Confidential and Proprietary

  20. Embedding Results Google Confidential and Proprietary

  21. Embedding Results Google Confidential and Proprietary

  22. Embedding Results Google Confidential and Proprietary

  23. Google Confidential and Proprietary

  24. Some Take Aways What Works ● ImageNet - of course! =) ● More data leads to better performance ● Deeper and bigger networks lead to better performance ● Networks handle many diverse problems very well What Needs Work ● More insight into the “Black Box” - diagnosis and understanding ● Understand and improve and training data efficiency ● Efficient means of collecting more training data ● Better ways to deal with noisy training data

  25. Thanks to the teams... ● Image Understanding Team ● Google Photos Team ● Google Brain Team ● Google Research ● Our great interns And We’re Hiring! I’m: chuck@google.com

  26. The End

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

EE-559 Deep learning 7. Networks for computer vision Fran cois Fleuret

EE-559 Deep learning 7. Networks for computer vision Fran cois Fleuret

EE-559 Deep learning 7. Networks for computer vision Fran cois Fleuret https://fleuret.org/dlc/ [version of: June 8, 2018] COLE POLYTECHNIQUE FDRALE DE LAUSANNE Tasks and data-sets Fran cois Fleuret EE-559 Deep learning

1.93k views • 170 slides
Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What

Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What

Neural Networks Hugo Larochelle ( @hugo_larochelle ) Google Brain 2 NEURAL NETWORKS What well cover ... f ( x ) computer vision architectures - convolutional networks - data augmentation 1 - residual networks ... ...

1.29k views • 75 slides
How Good is the Bayes Posterior in Deep Neural Networks Really? Florian Wenzel (Google Research

How Good is the Bayes Posterior in Deep Neural Networks Really? Florian Wenzel (Google Research

Code: github.com/google-research/google-research/tree/ master/cold_posterior_bnn How Good is the Bayes Posterior in Deep Neural Networks Really? Florian Wenzel (Google Research Berlin) Joint first authors: Kevin Roth, Bas Veeling, and: Jakub

385 views • 38 slides
Disentangling Trainability and Generalization In Deep Neural Networks Lechao Xiao, Je ff rey

Disentangling Trainability and Generalization In Deep Neural Networks Lechao Xiao, Je ff rey

Disentangling Trainability and Generalization In Deep Neural Networks Lechao Xiao, Je ff rey Pennington and Samuel S. Schoenholz Google Brain Team, Google Research Colab Tutorial Two Fundamental Theoretical Questions in Deep Learning

473 views • 15 slides
Deep Learning in Computer Vision Caner Hazrba Deep Learning in Action 24. June 15

Deep Learning in Computer Vision Caner Hazrba Deep Learning in Action 24. June 15

Deep Learning in Computer Vision Caner Hazrba Deep Learning in Action 24. June 15 Computer Vision Group 6 Postdocs, 16 PhD students Caner Hazrba | vision.in.tum.de Deep Learning in Computer Vision 2 Research in Computer Vision

960 views • 30 slides
Learning for Computer Vision Ramprasaath Lecture Outline Computer Vision Before

Learning for Computer Vision Ramprasaath Lecture Outline Computer Vision Before

A Shallow Introduction to Deep Learning for Computer Vision Ramprasaath Lecture Outline Computer Vision Before (Image/Alex)Net era (Summer 1956-2012) After (Image/Alex)Net era (2012 present) Neural Networks (Brief

1.2k views • 96 slides
Adversarial Training Attacks on Deep Networks and Generative Adversarial Networks Erkut Erdem

Adversarial Training Attacks on Deep Networks and Generative Adversarial Networks Erkut Erdem

images from Geris Game (Pixar, 1997) Adversarial Training Attacks on Deep Networks and Generative Adversarial Networks Erkut Erdem Aykut Erdem Levent Karacan Computer Vision Lab, Hacettepe University Outline Part 1: Attacks on

1.11k views • 72 slides
Deep multi-task learning with evolving weights Machine learning - computer vision published in

Deep multi-task learning with evolving weights Machine learning - computer vision published in

Deep multi-task learning with evolving weights Machine learning - computer vision published in European Symposium on Artificial Neural Networks (ESANN 2016) Soufiane Belharbi Romain Hrault Clment Chatelain Sbastien Adam

631 views • 44 slides
Delving Deep into Computer Vision Caner Hazirbas Machine Learning Meetup #1 Delving Deep into

Delving Deep into Computer Vision Caner Hazirbas Machine Learning Meetup #1 Delving Deep into

Delving Deep into Computer Vision Caner Hazirbas Machine Learning Meetup #1 Delving Deep into Computer Vision FlowNet FuseNet PoseLSTM DDFF Caner Hazirbas | hazirbas@cs.tum.edu Delving Deep into Computer Vision 2 Delving Deep into

1.14k views • 45 slides
Deep Convolutional Neural Network for Computer Vision Products LI XU, R&D Director

Deep Convolutional Neural Network for Computer Vision Products LI XU, R&D Director

Deep Convolutional Neural Network for Computer Vision Products LI XU, R&D Director SenseTime Group Limited SenseTime Introduction SenseTime focuses on invention and development of computer vision and deep learning technologies. Our

1.03k views • 39 slides
JOINT DETECTION AND SEGMENTATION WITH DEEP HIERARCHICAL NETWORKS Zhao Chen Machine Learning

JOINT DETECTION AND SEGMENTATION WITH DEEP HIERARCHICAL NETWORKS Zhao Chen Machine Learning

JOINT DETECTION AND SEGMENTATION WITH DEEP HIERARCHICAL NETWORKS Zhao Chen Machine Learning Intern, NVIDIA ABOUT ME 5th year PhD student in physics @ S tanford by day, deep learning computer vision scientist by night. Intern with Deep

1.1k views • 53 slides
Recurrent Neural Networks Deep neural networks have enabled major advances in machine learning

Recurrent Neural Networks Deep neural networks have enabled major advances in machine learning

Recurrent Neural Networks Deep neural networks have enabled major advances in machine learning and AI y t-1 y t y t+1 Computer vision h t-1 h t h t+1 Language translation Speech recognition h t-1 h t h t+1 Question answering x t-1 x t x t+1

454 views • 43 slides
Deep Neural Networks as Gaussian Processes Jaehoon Lee Google Brain Workshop on Accelerating the

Deep Neural Networks as Gaussian Processes Jaehoon Lee Google Brain Workshop on Accelerating the

Deep Neural Networks as Gaussian Processes Jaehoon Lee Google Brain Workshop on Accelerating the Search for Dark Matter with Machine Learning April 10, 2019 Based on Published in ICLR 2018, htups://arxiv.org/abs/1711.00165 Open source

501 views • 33 slides
Learning Visual Servoing with Deep Features and Fitted Q-Iteration Alex X. Lee 1 , 2,1,3 1 ,

Learning Visual Servoing with Deep Features and Fitted Q-Iteration Alex X. Lee 1 , 2,1,3 1 ,

Learning Visual Servoing with Deep Features and Fitted Q-Iteration Alex X. Lee 1 , 2,1,3 1 , Pieter Abbeel Sergey Levine 1 UC Berkeley, 2 OpenAI, 3 International Computer Science Institute Motivation Deep Neural Networks in Computer Vision

543 views • 35 slides
APPLICATIONS OF DEEP LEARNING TO COMPUTER VISION AND COMPUTER GRAPHICS Mike Houston Practical DEEP

APPLICATIONS OF DEEP LEARNING TO COMPUTER VISION AND COMPUTER GRAPHICS Mike Houston Practical DEEP

APPLICATIONS OF DEEP LEARNING TO COMPUTER VISION AND COMPUTER GRAPHICS Mike Houston Practical DEEP LEARNING Examples Image Classification, Object Detection, Localization, Speech Recognition, Speech Translation, Action Recognition, Scene

885 views • 40 slides
Deep Nets: What have they ever done for Vision? Alan Yuille Dept. Cognitive Science and

Deep Nets: What have they ever done for Vision? Alan Yuille Dept. Cognitive Science and

Deep Nets: What have they ever done for Vision? Alan Yuille Dept. Cognitive Science and Computer Science Johns Hopkins University What Have Deep Nets done to Computer Vision? Compared to human observers, Deep Nets are brittle and rely

460 views • 26 slides
A recent reversal of climatic trends in the North Atlantic Jon Robson j.i.robson@reading.ac.uk

A recent reversal of climatic trends in the North Atlantic Jon Robson j.i.robson@reading.ac.uk

A recent reversal of climatic trends in the North Atlantic Jon Robson j.i.robson@reading.ac.uk @JonIRobson Thanks to : Pablo Ortega and Rowan Sutton Motivation: Atlantic Multi-decadal Variability Robson et al, 2012 Substantial evidence that

681 views • 25 slides
You cannot make bricks without clay Cenex, Vianova & SRM Reti e Mobilit (Bologna) 2

You cannot make bricks without clay Cenex, Vianova & SRM Reti e Mobilit (Bologna) 2

You cannot make bricks without clay Cenex, Vianova & SRM Reti e Mobilit (Bologna) 2 Data! Data! Data! You cannot make bricks he e cri ried ed impatien patiently. tly. without clay. I cannot make bricks without clay.

1.05k views • 73 slides
Who We Are Our mission at CFED is to make it possible for

Who We Are Our mission at CFED is to make it possible for

Who We Are Our mission at CFED is to make it possible for millions of people to achieve financial security and contribute to an opportunity economy. UW-Extension Financial Coaching

826 views • 78 slides
RELATIVE SET THEORY Karel Hrbacek Department of Mathematics The City College of New York This is

RELATIVE SET THEORY Karel Hrbacek Department of Mathematics The City College of New York This is

RELATIVE SET THEORY Karel Hrbacek Department of Mathematics The City College of New York This is a report on a work in progress . Partial results are in (1) Internally iterated ultrapowers , in: Nonstandard Models of Arithmetic and Set Theory ,

450 views • 20 slides
Unifying Adjunct Islands and Freezing Effects in Minimalist Grammars Tim Hunter Department of

Unifying Adjunct Islands and Freezing Effects in Minimalist Grammars Tim Hunter Department of

Overview of Stabler 2006 Adding an Implementation of Adjunction Empirical Payoff Conclusion Unifying Adjunct Islands and Freezing Effects in Minimalist Grammars Tim Hunter Department of Linguistics University of Maryland TAG+10 1 / 26

948 views • 65 slides
Renewal Event September 17 th , 2015 Meeting Objectives 1. Renew

Renewal Event September 17 th , 2015 Meeting Objectives 1. Renew

Renewal Event September 17 th , 2015 Meeting Objectives 1. Renew our joint resolution 2. Gain strategic guidance from senior leaders 3. Review nature and

650 views • 29 slides
Quality Improvement for NHS Board Members Monday 26 th February 2018 Event Code #QIBM SRU guest

Quality Improvement for NHS Board Members Monday 26 th February 2018 Event Code #QIBM SRU guest

Quality Improvement for NHS Board Members Monday 26 th February 2018 Event Code #QIBM SRU guest *no password* #qiforboardmembers Welcome David Garbutt Chair of NHS Chairs Quality Portfolio Group and Scottish Ambulance Service

1.77k views • 128 slides
1 - BUNDLE PRODUCT What do we mean. Some History A pipeline bundle is a package of subsea

1 - BUNDLE PRODUCT What do we mean. Some History A pipeline bundle is a package of subsea

S afe C lean S mart F air A nywhere 1 - BUNDLE PRODUCT What do we mean. Some History A pipeline bundle is a package of subsea well or step out field development facilities. Typically pipelines control umbilicals and manifolds. The Manifold

857 views • 69 slides

More recommend