Adversarial Networks : When fake never looked so real Evan Ntavelis - PowerPoint PPT Presentation

Apr 06, 2024 •281 likes •475 views

Generative Adversarial Networks : When fake never looked so real Evan Ntavelis 1,2 Dr. Iason Kastanis 1 Philipp Schmid 1 {ens, iks, psd}@csem.ch 1. Robotics & Machine Learning CSEM SA 2. Computer Vision Lab ETH Zrich CSEM at a glance

Generative Adversarial Networks : When fake never looked so real Evan Ntavelis 1,2 Dr. Iason Kastanis 1 Philipp Schmid 1 {ens, iks, psd}@csem.ch 1. Robotics & Machine Learning CSEM SA 2. Computer Vision Lab ETH Zürich
CSEM at a glance – Close to industry Zürich Alpnach Muttenz M Z Neuchâtel Landquart N L A 2 83.0 450 175 64 Turnover Persons Industrial European (mio CHF) clients projects
Technologies in focus at CSEM 3
4 Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks Zhu et al. 2017
5 AttnGAN: Fine-Grained Text to Image Generation with Attentional Generative Adversarial Networks} Xu et al 2018
6 A Style-Based Generator Architecture for Generative Adversarial Networks Karras et al. 2018
7 Semantic Image Synthesis with Spatially-Adaptive Normalization Park et al. 2019
8 Source: datagrid.co.jp 2019
9 Few-Shot Adversarial Learning of Realistic Neural Talking Head Models Zakharov et al. 2019
Generative Adversarial Nets • Introduced in 2014 by Ian Goodfellow 10 • Rapidly Adopted • Unprecedented Generational Quality
Generative Adversarial Nets • An adversarial game between two subnets: 11 • The Generator • The Discriminator
Deep Fakes • In the era of Fake News do highly realistic images harbor dangers to the society? 12
Defense Mechanisms 13
The important question… 14 How can we use GANs in the industry?
The Problem • Gathering data is tedious and costly • Good quality labels require even more effort 15
A Solution Using Adversarial Networks • Adversarial Domain Adaptation • Train on a simulated data and adapt for the use case • Data Augmentation 16 • Learn how to generate new samples to train with • Generate images with desired attributes Sources: CyCADA: Cycle-Consistent Adversarial Domain Adaptation Hoffman et al. 2017, GAN-based Synthetic Medical Image Augmentation for increased CNN Performance in Liver Lesion Classification Frid-Adar et al, 2018
But… • GANs are not a panacea • Nascent technology • Difficult to train • Require abundance of data 17 • Clever schemes may reduce the effort • Yet, very promising results • Worth the effort!
That’s all folks! Are you interested in being part of a highly stimulating environment working on the latest Deep Learning Technologies? 18 We are hiring!

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

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

Generative Adversarial Networks, Wasserstein Distance, and Adversarial Loss Zhiyu Min Alibaba

Generative Adversarial Networks, Wasserstein Distance, and Adversarial Loss Zhiyu Min Alibaba AliMe X-Lab Outline GAN Definition and formulation Saddle point optimization Vanishing gradient Alternative objective for Generator

516 views • 25 slides

Neglected topics CS 446 Adversarial examples and deep networks 1 / 23 Adversarial

Neglected topics CS 446 Adversarial examples and deep networks 1 / 23 Adversarial examples? Standard ML setup: We have training data; try to do well on withheld testing data. Adversarial/robust ML setup: We have training

672 views • 27 slides

LAB MEETING: A Connection Between Generative Adversarial Networks, Inverse Reinforcement Learning

LAB MEETING: A Connection Between Generative Adversarial Networks, Inverse Reinforcement Learning and Energy-Based models Suwon Suh POSTECH MLG Feb, 13, 2017 Goal Understanding Basic Models 1) Generative Adversarial Networks (GAN) 2)

722 views • 12 slides

Introduction to Generative Adversarial Networks Ian Goodfellow, OpenAI Research Scientist NIPS

Introduction to Generative Adversarial Networks Ian Goodfellow, OpenAI Research Scientist NIPS 2016 Workshop on Adversarial Training Barcelona, 2016-12-9 Adversarial Training A phrase whose usage is in flux; a new term that applies to both

792 views • 25 slides

Robust Statistics and Generative Adversarial Networks Yuan YAO HKUST Chao Gao (Chicago) Jiyu

Robust Statistics and Generative Adversarial Networks Yuan YAO HKUST Chao Gao (Chicago) Jiyu Liu (Yale) Weizhi Zhu (HKUST) Deep Learning is Notoriously Not Robust! Imperceivable adversarial examples are ubiquitous to fail neural networks

1.8k views • 133 slides

Generative Adversarial Networks presented by Ian Goodfellow presentation co-developed with Aaron

Generative Adversarial Networks presented by Ian Goodfellow presentation co-developed with Aaron Courville 1 In todays talk Generative Adversarial Networks Goodfellow et al., NIPS 2014 Conditional Generative

2.84k views • 40 slides

Genera&ve Adversarial Networks NTT

Genera&ve Adversarial Networks NTT 2020 6 10 @JSAI2020 KS-02 AI MIRU2020

899 views • 70 slides

Generative Adversarial Networks Aaron Mishkin UBC MLRG 2018W2 1 Generative Adversial Networks

Generative Adversarial Networks Aaron Mishkin UBC MLRG 2018W2 1 Generative Adversial Networks Two imaginary celebrities that were dreamed up by a random number generator. https://research.nvidia.com/publication/2017-10

665 views • 39 slides

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist Security

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist Security Seminar, Stanford University, 2017-01-17 In this presentation Intriguing Properties of Neural Networks Szegedy et al, 2013

939 views • 44 slides

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist Guest

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist Guest lecture for CS 294-131, UC Berkeley, 2016-10-05 In this presentation Intriguing Properties of Neural Networks Szegedy et al, 2013

660 views • 44 slides

Robust Statistics and Generative Adversarial Networks Yuan YAO HKUST 1 Chao GAO Jiyi LIU

Robust Statistics and Generative Adversarial Networks Yuan YAO HKUST 1 Chao GAO Jiyi LIU Weizhi ZHU U. Chicago Yale U. HKUST 2 Deep Learning is Notoriously Not Robust! Imperceivable adversarial examples are ubiquitous to fail

833 views • 72 slides

CVPR 2020 Universal Adversarial Attacks Image agnostic and transferable across networks

1 1,4 2,3 2 3 4 1 CVPR 2020 Universal Adversarial Attacks Image agnostic and transferable across networks Adversarial No defense: Baseline DNN perturbation Input image Classification Feature extraction dense Perturbed image

592 views • 6 slides

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist

Adversarial Examples and Adversarial Training Ian Goodfellow, OpenAI Research Scientist Presentation at San Francisco AI Meetup, 2016-08-18 In this presentation Intriguing Properties of Neural Networks Szegedy et al, 2013

690 views • 32 slides

ON THE ADVERSARIAL ROBUSTNESS OF UNCERTAINTY AWARE DEEP NEURAL NETWORKS APRIL 29 TH , 2019

ON THE ADVERSARIAL ROBUSTNESS OF UNCERTAINTY AWARE DEEP NEURAL NETWORKS APRIL 29 TH , 2019 PREPARED BY: ALI HARAKEH QUESTION Can a neural network mitigate the effects of adversarial attacks by estimating the uncertainty in its predictions ?

1.14k views • 26 slides

Generative Adversarial Networks Wanyue Li

Generating Fundus Fluorescence Angiography Images from Structure Fundus Images Using Generative Adversarial Networks Wanyue Li (wanyueli93@126.com) University of Science and Technology of China (USTC)

144 views • 10 slides

Imaging biomarkers in oncologic liver disease Bernard Van Beers Laboratory of Imaging Biomarkers

Imaging biomarkers in oncologic liver disease Bernard Van Beers Laboratory of Imaging Biomarkers INSERM UMR1149 University Paris Diderot Department of Radiology Beaujon University Hospital Paris Nord Imaging biomarkers Imaging

711 views • 25 slides

The BIST History of FPGAs FPGAs The BIST History of The BISTory BISTory of of FPGAs FPGAs

The BIST History of FPGAs FPGAs The BIST History of The BISTory BISTory of of FPGAs FPGAs The Chuck Stroud Chuck Stroud Electrical and Computer Engineering Electrical and Computer Engineering Auburn University Auburn University 1

685 views • 42 slides

Creating transcription helper f u nctions SP OK E N L AN G U AG E P R OC E SSIN G IN P YTH ON

Creating transcription helper f u nctions SP OK E N L AN G U AG E P R OC E SSIN G IN P YTH ON Daniel Bo u rke Machine Learning Engineer / Yo u T u be Creator E x ploring a u dio files # Import os module import os # Check the folder of

816 views • 46 slides

Silicon PUFs and PUF-based Key Storage Roel Maes Intrinsic-ID, Eindhoven (NL) June 6, 2014

Silicon PUFs and PUF-based Key Storage Roel Maes Intrinsic-ID, Eindhoven (NL) June 6, 2014 Summerschool: Design and security of cryptographic algorithms and devices for real-world applications ibenik , Croatia Roots of Trust Entity Data

749 views • 28 slides

Hydatid cyst of the liver treated with Albendazole: (a) After 1 cycle (3 month); (b) 6 months

imaging methods in the diagnosis of hydatid cyst disease Fatemeh mohammadi chelkasari Radiologist MD Shahid rajaei hospital Hydatid cyst of the liver treated with Albendazole: (a) After 1 cycle (3 month); (b) 6

651 views • 25 slides

COMP 135 Introduction to Machine Learning First day of class Spring 2019

COMP 135 Introduction to Machine Learning First day of class Spring 2019 https://www.cs.tufts.edu/comp/135/2019s/ Many slides attributable to: Emily Fox (UW), Finale Doshi-Velez (Harvard), Erik Sudderth (UCI), & Liping Liu (Tufts) 1

497 views • 38 slides

For others that inspire A Love That Paid The Price (Mark 14:32 - 42 We all understand that if we

For others that inspire A Love That Paid The Price (Mark 14:32 - 42 We all understand that if we owe debt to someone, we must pay our debts. God out of His love for suffering humanity offered His Son Jesus Christ to pay our sin debt in full,

540 views • 27 slides

Chines nese A Amer erica can H Hem ematologist Onco Oncologist Net Netwo work PROUDLY

Chines nese A Amer erica can H Hem ematologist Onco Oncologist Net Netwo work PROUDLY RECOGNIZE Yanfeng Liu, Yijun Yang, Xiaojin Sha, Ningfei An AS THE WINNERS OF THE 2012 CAHON YOUNG INVESTIGATOR AWARD __Signature_________________

650 views • 5 slides