Ramakrishna

@heremytatersat

It's happening!

Vrijeme pridruživanja: srpanj 2014.
3 fotografije ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @heremytatersat

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @heremytatersat

  1. proslijedio/la je Tweet
    31. sij

    A thread with a list of people who have few followers (I put an arbitrary cut at 2k) but regularly post interesting stuff about science (mostly math and physics, as I am biased toward these disciplines 😉) In strictly random order. (1/n)

    5 replies 55 proslijeđenih tweetova 244 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    13. sij

    An exciting new paper argues that mainstream economic theory has made a mistake in how it models individuals making risky decisions, and offers a mathematical correction which takes the passage of time into account.

    1 reply 29 proslijeđenih tweetova 62 korisnika označavaju da im se sviđa
    Poništi
  3. proslijedio/la je Tweet
    28. stu 2019.

    High time to read: "Causality for Machine Learning" by Bernhard Schölkopf

    1 reply 3 proslijeđena tweeta 13 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    28. stu 2019.

    Excited to share our work on Contrastive Learning of Structured World Models! C-SWMs learn object-factorized models & discover objects without supervision, using a simple loss inspired by work on graph embeddings Paper: Code: 1/5

    137 proslijeđenih tweetova 523 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    22. stu 2019.

    Generating Interactive Worlds with Text @ AAAI'20 Built in ParlAI Angela Fan Pratik Ringshia Emma Qian Arthur Szlam

    43 proslijeđena tweeta 154 korisnika označavaju da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    21. lis 2018.
    Odgovor korisnicima

    There are a couple questions in that ethics section you can ask every time and they always take similar form, such as bias w.r.t. predefined groups of users. How do we prevent that section becoming a box-ticking excercise and allow people to really think about what can go wrong?

    1 reply 1 proslijeđeni tweet 1 korisnik označava da mu se sviđa
    Poništi
  7. proslijedio/la je Tweet
    16. stu 2019.

    a proposal: people-centric language but for algorithms. ❌ the algorithm kicked people off benefits ✅ people in power used an algorithm to kick people off benefits ❌ the algorithm cleared criminal convictions ✅ people in power used an algorithm to clear criminal convictions

    21 reply 2.698 proslijeđenih tweetova 7.806 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    16. stu 2019.

    You can't just defer to data. You have to keep inspecting your results, looking for feedback loops, iterating, and visualizing in order to identify & mitigate risks. -- Deven Desai

    1 reply 9 proslijeđenih tweetova 53 korisnika označavaju da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    13. stu 2019.

    "Learning From [Mouse] Brains How to Regularize Machines," Li et al.: Wild stuff - they showed images to mice, recorded the mice's neural activity, made a model of that, then penalized not-mouse-brain-y representations when training new classifiers.

    64 proslijeđena tweeta 226 korisnika označava da im se sviđa
    Poništi
  10. proslijedio/la je Tweet
    12. stu 2018.

    latent weavings at the edge of the mode collapse

    1 proslijeđeni tweet 10 korisnika označava da im se sviđa
    Poništi
  11. proslijedio/la je Tweet
    10. stu 2019.

    Medicine is only effective because biology has self-repair capabilities. If our biology were built like our technology then our medical practices would be unable to overcome the complexity.

    1 reply 2 proslijeđena tweeta 4 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    9. stu 2019.

    States force the world into simplistic models to make it legible to bureaucracy, sometimes at the cost of great harm. (see James Scott) Software often forces even more simple and rigid schemas. But deep learning makes fuzzy human ideas computable. Can it reverse the trend?

    14 replies 51 proslijeđeni tweet 210 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    7. stu 2019.

    1/In today's post I yell at Lant Pritchett for yelling at Banerjee, Duflo and Kremer:

    10 replies 109 proslijeđenih tweetova 252 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    27. svi 2019.

    The first six books of the Elements of Euclid, typeset using TeX by

    7 replies 173 proslijeđena tweeta 552 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    5. stu 2019.

    When regularizing « optimal transport like » problems, you should use the relative entropy (aka mutual information) and not the discrete entropy. Normalizing your entropy is the key to success!

    72 proslijeđena tweeta 347 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    4. stu 2019.

    One under appreciated fact highlighted in this article: the theoretical receptive field of modern neural networks is *ridiculously* large. Often thousands of pixels, while the input image ins only hundreds.

    Computing Receptive Fields of Convolutional Neural Networks -- A new Distill article by André Araujo, Wade Norris, and Jack Sim.
    Prikaži ovu nit
    36 proslijeđenih tweetova 205 korisnika označava da im se sviđa
    Poništi
  17. proslijedio/la je Tweet
    23. lis 2019.

    Online version? has you covered: (AND IT’S THE FEATURED ARTICLE ON THEIR HOMEPAGE 😭)

    4 proslijeđena tweeta 14 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    1. stu 2019.

    Our gallery is now live on Congrats to and all the other short and longlisted artists 🎉🤖 Organised together with

    Korisnici , , i
    1 reply 18 proslijeđenih tweetova 57 korisnika označava da im se sviđa
    Poništi
  19. proslijedio/la je Tweet
    19. velj 2019.

    whenever we have a disagreement, the first thing I want to understand is: how have your experiences have been different from mine?

    8 replies 25 proslijeđenih tweetova 102 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    1. stu 2019.

    Excited to share 's work (with minor contributions by myself) on the Generative Graph Transformer, to be presented at the Graph Representation Learning workshop at . Blog: Paper: 1/2

    1 reply 53 proslijeđena tweeta 225 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi

@heremytatersat još nije objavio nijedan Tweet.

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·