UGOTAG | Videos with Chapter Markers  
  
× ANCIENT WORLD ANIMALS AQUARIUM ARCHERY ARCHITECTURE ART ARTIFICIAL INTELLIGENCE AUDIO BOOKS AVIATION BABY BEAUTY BIG THOUGHTS BIKING BIOLOGY BIRDS BLOCKCHAIN BUSINESS AUTOMOTIVE CATS CHRISTMAS COLD WAR COMPUTER SECURITY CRAFTS CRIME STORIES CRYPTOCURRENCY CSGO DIGITAL COMBAT SIMULATOR DINOSAUR DOGS EASTER ECONOMICS ELECTRONICS ENGINEERING ENVIRONMENT FAR CRY 5 FARMING FASHION FISHING FITNESS FOOD & DRINK FORTNITE BATTLE ROYALE FREEDOM OF SPEECH FUNNY GAMBLING GAMING GARDENING GEOPOLITICS GOD OF WAR GUITAR GUNS HALLOWEEN HARRY POTTER HEALTH & WELLNESS HISTORY YOUR HOME HOME REPAIR HOMEBREW INSECTS INTERESTING KIDS KITCHEN KNITTING LAWNCARE LEGO LIQUOR LOCKSPORT MARKETING MARTIAL ARTS MATH MENS STYLE MINECRAFT MOBILE DEVICES MOTORBIKES MOVIES MULTIPLICATION MUSIC MYSTERY NEUROSCIENCE OLYMPICS ORGANIZATION OUTDOORS PETS PHILOSOPHY PHOTOGRAPHY PHYSICS PI DAY POLITICS POTTERY PRIVACY PROGRAMMING PSYCHOLOGY RECIPE RED DEAD REDEMPTION RELAXING RELIGION REMOTE CONTROL ROCKCLIMBING SCI FI SCI FY SCIENCE SHOOTING SPORTS SKATEBOARDING SPECIAL FORCES SPEECHES SPORTS STAR WARS STEM STPATRICKS STYLE TECHNOLOGY THANKSGIVING TOYS TRAVEL TV VALENTINE'S DAY WAR WEDDING WOODWORKING WW1 WORLD WAR 2
HOME  |  TECHNOLOGY  |  HEALTH  |  FOOD  |  MORE

Episode Markers
  • 03:41
     
    #feature engineering for free   
    With deep learning, you get feature engineering for free. At least with some caveats.
  • 04:40
     
    #one infrastructure for multiple problems   
    Deep learning promises one infrastructure for multiple problems.
  • 05:35
     
    #hierarchical learning   
    Deep learning provides for hierarchical learning.
  • 05:59
     
    #efficient and easily parallelized   
    Deep learning is efficient and easily parallelized.
  • 08:22
     
    #How training works in deep neural networks   
    How training works in deep neural networks. Step 1 -- Feed Forward.
  • 09:36
     
    #Backpropagation   
    Step 2 -- Backpropagation . Assign blame, and feed the error back in the network.
  • 11:02
     
    #Stochastic Gradient Descent   
    This optimization is done via algorithms like Stochastic Gradient Descent which finds the local minima in a problem space.
  • 13:45
     
    #Demo of Google's TensorFlow   
    Demo of Google's TensorFlow deep learning framework and the MNIST handwritten digits database.
  • 17:20
     
    #find spammy images   
    One way to find spammy images is to pass them through a convolution neural network, and extract the second layer features. These second layer features can be rapidly analyzed by a human for best results, effectively creating a cyborg detection system. This is used by Facebook.
  • 27:00
     
    #Adversarial training of a deep learning model   
    Adversarial training of a deep learning model using cheque images.
  • 37:34
     
    #captcha cruncher   
    captcha cruncher

Adversarial Attacks on Deep Learning

Deep learning and neural networks have gained incredible popularity in recent years. The technology has grown to be the most talked-about and least well-understood branch of machine learning. Aside from it�s highly publicized victories in playing Go, numerous successful applications of deep learning in image and speech recognition has kickstarted movements to integrate it into critical fields like medical imaging and self-driving cars. In the security field, deep learning has shown good experimental results in malware/anomaly detection, APT protection, spam/phishing detection, and traffic identification. This DEF CON 101 session will guide the audience through the theory and motivations behind deep learning systems. We look at the simplest form of neural networks, then explore how variations such as convolutional neural networks and recurrent neural networks can be used to solve real problems with an unreasonable effectiveness. Then, we demonstrate that most deep learning systems are n






Community tags: computer_security     HOME     SIGN UP     CONTACT US