pinboard July 14, 2017

• liufuyang/kaggle-youtube-8m
• [1701.00160] NIPS 2016 Tutorial: Generative Adversarial Networks
  This report summarizes the tutorial presented by the author at NIPS 2016 on generative adversarial networks (GANs). The tutorial describes: (1) Why generative modeling is a topic worth studying, (2) how generative models work, and how GANs compare to other generative models, (3) the details of how GANs work, (4) research frontiers in GANs, and (5) state-of-the-art image models that combine GANs with other methods. Finally, the tutorial contains three exercises for readers to complete, and the solutions to these exercises.
• GV / Team
• [1609.04802] Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
  Despite the breakthroughs in accuracy and speed of single image super-resolution using faster and deeper convolutional neural networks, one central problem remains largely unsolved: how do we recover the finer texture details when we super-resolve at large upscaling factors? The behavior of optimization-based super-resolution methods is principally driven by the choice of the objective function. Recent work has largely focused on minimizing the mean squared reconstruction error. The resulting estimates have high peak signal-to-noise ratios, but they are often lacking high-frequency details and are perceptually unsatisfying in the sense that they fail to match the fidelity expected at the higher resolution. In this paper, we present SRGAN, a generative adversarial network (GAN) for image super-resolution (SR). To our knowledge, it is the first framework capable of inferring photo-realistic natural images for 4x upscaling factors. To achieve this, we propose a perceptual loss function which consists of an adversarial loss and a content loss. The adversarial loss pushes our solution to the natural image manifold using a discriminator network that is trained to differentiate between the super-resolved images and original photo-realistic images. In addition, we use a content loss motivated by perceptual similarity instead of similarity in pixel space. Our deep residual network is able to recover photo-realistic textures from heavily downsampled images on public benchmarks. An extensive mean-opinion-score (MOS) test shows hugely significant gains in perceptual quality using SRGAN. The MOS scores obtained with SRGAN are closer to those of the original high-resolution images than to those obtained with any state-of-the-art method.
• Overview
  Google’s Community Space
• Twitter
  RT @doctorow: engineer Karen Leadlay working on the analog computers in the space division of General Dynamics, 1964.…
• Twitter
  RT @kbeninato: Maybe it sounded more legit in the original Russian.
• Research Blog: Federated Learning: Collaborative Machine Learning without Centralized Training Data
  Standard machine learning approaches require centralizing the training data
  on one machine or in a datacenter. And Google has built one of the most
  secure and robust cloud infrastructures for processing this data to make
  our services better. Now for models trained from user interaction with
  mobile devices, we’re introducing an additional approach: *Federated
  Learning*.
  Federated Learning enables mobile phones to collaboratively
  learn a shared prediction model while keeping all the training data on
  device, decoupling the ability to do machine learning from the need to
  store the data in the cloud. This goes beyond the use of local models that
  make predictions on mobile devices (like the Mobile Vision API
  <https://developers.google.com/vision/> and On-Device Smart Reply
  <https://research.googleblog.com/2017/02/on-device-machine-intelligence.html>)
  by bringing model *training* to the device as well.
• Twitter
  RT @modestproposal1: Holy shit the transcript is real he really said the wall needs to be transparent so you don’t get hit in the head b…
• Twitter
  RT @PhilipPullman: Absolutely extraordinary. How is this happening? Why can’t anyone stop it?
• Excerpts From Trump’s Conversation With Journalists on Air Force One – NYTimes.com
• Twitter
  RT @RichardWiseman: The amazing @PaulNoth says more about religion in 1 cartoon than academics in 100 books
• Twitter
  RT @JamesOMCraig: Today’s #graffitioftheday.
• Twitter
  RT @michikokakutani: Auvers Town Hall in July. by Vincent van Gogh. 1890.
• With Glare on Trump Children, Political Gets Personal for President – The New York Times
• Blow book – Wikipedia
• Twitter
  RT @myblackmindd: A guy was projecting these images onto a wall in downtown Oakland just now.
• Twitter
  RT @BeschlossDC: San Francisco at night, 1961: #Wiener
• The key to the Trump-Russia scandal? Follow the data

  What’s bigger than Don Jr.’s email? If Trump campaign, Mercer firm and Russia colluded on fake news. Here’s the case http://www.philly.com/philly/blogs/attytood/the-key-to-the-trump-russia-scandal-follow-the-data-20170713.html

• Rocket Prose — Classic animators doing reference poses for their…
• A Grand Unified Theory of Avocado Toast | The New Yorker
• News | NASA’s Juno Spacecraft Spots Jupiter’s Great Red Spot
• Untitled (http://www.theroot.com/black-teenager-mistaken-for-larger-bald-black-man-says-1796840022)
  RT @iSmashFizzle: Cops claim they thought a 5’2 115 pound black girl was a 5’10 170 pound black man. And that’s why they punched her:
• Perform sentiment analysis with LSTMs, using TensorFlow – O’Reilly Media
• Google partners with VCs to host its own machine learning startup competition | TechCrunch
• Here are the winners of the Google Cloud machine learning pitch-off | TechCrunch
• How to do text classification with CNNs, TensorFlow and word embedding
• The Making of ‘Meatballs’: Is Bill Murray Even Going to Show Up?
• Twitter
  RT @ftrain: I feel that this was a good tweet for 2014.

Digest powered by RSS Digest