Gangadhar Payyavula

@Gangadhar_P

Software engineer at Amazon. works with large datasets and also dabble in deep learning research.

Seattle, WA
Joined October 2008
0 Photos and videos Photos and videos

Tweets

You blocked @Gangadhar_P

Are you sure you want to view these Tweets? Viewing Tweets won't unblock @Gangadhar_P

  1. Retweeted
    Jun 9

    Machine learning and deep learning are 99% open source. The competitive advantage is in training data.

    13 replies 221 retweets 382 likes
  2. Retweeted
    Jun 9

    The "Whoa are you serious" award for an Appendix goes to "Self-Normalizing Neural Networks" proposes "selu" nonlin

    16 replies 107 retweets 339 likes
  3. Retweeted
    Jun 9

    Handcrafted, bespoke, one-of-a-kind activation function. As a neural network artisan, I can't wait to try it

    Replying to @hardmaru
    This is the magic SELU activation function.
    7 replies 65 retweets 245 likes
  4. Retweeted
    Jun 6

    Tupper's self-referential formula is a formula that visually represents itself when graphed at a specific location in the (x, y) plane.

    43 replies 1,467 retweets 2,569 likes
  5. Retweeted
    Jun 6

    Visual Interaction Networks: Can learn to predict 100s of future physical states from just a few frames of video.

    1 reply 46 retweets 112 likes
  6. Retweeted
    Jun 6

    Cool architecture for relational reasoning by et al!

    1 reply 62 retweets 184 likes
  7. Retweeted
    Jun 6

    "Data is oil" is old saying.."Training data is new oil" is new saying.

    9 retweets 10 likes
  8. Retweeted
    Jun 5

    Notes on two recent papers on embeddings. Addresses better mapping for latent hierarchies and embedding rare words.

    17 retweets 69 likes
  9. Retweeted
    Jun 5

    One of the most important deep learning papers of the year, thus far.

    Excited about our new paper on relational reasoning, with and others at DeepMind.
    4 replies 116 retweets 336 likes
  10. Retweeted
    Jun 4

    This overview table is used by the team to decide which visualization to use. It's a handy starting point. HT

    , , and 7 others
    20 replies 753 retweets 1,536 likes
  11. Retweeted
    Jun 4

    Our new paper, "PixelGAN autoencoders": semi-supervised learning with GAN inference networks and PixelCNN decoders

    37 retweets 123 likes
  12. Retweeted
    May 25

    My paper with Joan Bruna: graph neural networks applied to community detection/clustering!

    3 replies 27 retweets 89 likes
  13. Retweeted
    May 14

    was started at Princeton but completed at Stanford, a genuine collaboration of multi universities.

    Replying to @tiqerlu
    People please take note— is right. ImageNet was done by while she was at Princeton, not Stanford.
    3 replies 19 retweets 99 likes
  14. Retweeted
    Jun 3

    In design, paying keen attention to the little details matter a lot. A detailed analysis of google and apple's maps.

    5 replies 5 retweets 61 likes
  15. Retweeted
    Jun 1

    Wasserstein GAN is dead, long live the Crammer GAN!

    1 reply 69 retweets 189 likes
  16. Retweeted
    May 31

    Our new paper is out! Surprising results re: Wasserstein distance. Expecting plenty of Seinfeld jokes. W/ and

    The Cramer Distance as a Solution to Biased Wasserstein Gradients. (arXiv:1705.10743v1 [cs.AI])
    2 replies 24 retweets 64 likes
  17. Retweeted
    May 30

    Good Semi-supervised Learning that Requires a Bad GAN.

    21 retweets 42 likes
  18. Retweeted
    Apr 8

    Probably the most epic neural network visualization ever...

    6 replies 200 retweets 310 likes
  19. Retweeted
    May 28

    Another paper on regularizing gradients for GANs (see also: VEEGAN, WGAN-GP). All frameworks should (correctly) support grads of grads!

    Stabilizing Training of Generative Adversarial Networks through Regularization. (arXiv:1705.09367v1 [cs.LG])
    1 reply 27 retweets 88 likes
  20. Retweeted
    May 26
    Replying to

    When you're building AI-powered interfaces, the use cases for a dynamically generated frontend come up surprisingly often!

    1 reply 2 retweets 5 likes

@Gangadhar_P hasn't Tweeted yet.

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

    You may also like

    ·