hi o

@_hi_o_

Vrijeme pridruživanja: travanj 2018.
1 fotografija ili videozapis Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @_hi_o_

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

  1. proslijedio/la je Tweet
    4. velj

    1 reply 50 proslijeđenih tweetova 391 korisnik označava da mu se sviđa
    Poništi
  2. proslijedio/la je Tweet
    2. velj

    Portugal's government now wants to restrict "golden visa" incentives for foreign property buyers to purchases outside Lisbon and the northern city of Oporto via

    1 reply 22 proslijeđena tweeta 41 korisnik označava da mu se sviđa
    Poništi
  3. proslijedio/la je Tweet
    2. velj

    For perspective, since China first reported it Dec. 30, the coronavirus has killed 305 people worldwide while car crashes have killed about 125,000.

    At least 305 people have died and confirmed cases surged to more than 14,500 globally. We're tracking the spread of the coronavirus
    57 replies 731 proslijeđeni tweet 2.198 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    31. sij

    Success: I trained ResNet-50 on imagenet to 75.9% top-1 accuracy in 3.51 minutes using a 512-core TPUv3. (480,000 images per second. 224x224 res JPG.) Before you think highly of me, all I did was run Google’s code. It was hard though. Logs:

    37 proslijeđenih tweetova 200 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    1. velj

    A great way to realise the sheer quantity of information our 🧠 seamlessly processes when we do a “simple” task like driving a car. This is what the Tesla autopilot sees using a neural network which took 70,000 GPU hours to train. ht

    499 proslijeđenih tweetova 1.266 korisnika označava da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    25. sij

    +1! IMHO the most successful machine learning system of all, internet search, is a great example of this. search is just a super powered externalised memory. the next-big-thing (TM) will be super powered externalised cognition. we want to augment ppl rather than automating them.

    Honestly, the question is not, and has never been, "can machine learning replace radiologists/etc" (which won't happen in the foreseeable future). The question is, how can radiology/etc utilize ML to improve outcomes, decrease the cost of care, and broaden accessibility.
    Prikaži ovu nit
    14 proslijeđenih tweetova 88 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    23. sij

    : Confirmed - A new hospital in in is being build right now, and will be finished within 6 days, to transfer all patients with the only symptoms of the to begin cleaning all of the other hospitals. The hospital will be demolished afterwards.

    114 replies 1.853 proslijeđena tweeta 2.064 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    5. sij

    Lloyd’s algorithm is the continuous counterpart of k-means. Optimizes the optimal quantization energy, which is an optimal transport distance to free Dirac masses.

    11 replies 301 proslijeđeni tweet 1.445 korisnika označava da im se sviđa
    Poništi
  9. proslijedio/la je Tweet
    5. sij

    The argument for VAE disentangling representations is that the latent state is normal hence its components are independent. Any rotation would entangle without changing the distribution, so why have components often a clear semantic?

    11 proslijeđenih tweetova 39 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    23. sij

    A mixture of Gaussians model can be represented as a weighted points cloud (actually a measure) over the mean/covariance domain. Mixture fitting is a non-convex optimization problem.

    6 replies 110 proslijeđenih tweetova 469 korisnika označava da im se sviđa
    Poništi
  11. proslijedio/la je Tweet
    19. sij

    Had to verify. And yes. Kernighan and Ritchie really did this. TIL :)

    101 reply 2.712 proslijeđenih tweetova 8.310 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    3. sij

    A General and Adaptive Robust Loss Function They propose an analytical function that can represent a family of well known robust cost functions just with a single parameter (alpha). Alpha lets you walk through L2, huber, cauchy, tukey and more.

    14 replies 276 proslijeđenih tweetova 1.066 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    3. sij

    I could propose such predictor: until GANs are not used on , they don't help.

    "Seeing is Not Necessarily Believing: Limitations of BigGANs for Data Augmentation" Tl;dr: "ImageNet train on only BigGAN images:,Top-1 err incr 120%, top5 -384%. Mixing with real <= real only." No realible predictor if GAN could help by
    1 proslijeđeni tweet 10 korisnika označava da im se sviđa
    Poništi
  14. proslijedio/la je Tweet
    16. pro 2019.

    Before ML Twitter gets sick of hearing about , I wanted to highlight a really cool talk by Prof. Zhiru Zhang from Cornell about Neural Network and Hardware co-design. Here’s a great summary slide giving an overview of different NN model optimization techniques!

    9 proslijeđenih tweetova 67 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    16. pro 2019.

    Solving linear systems comes with different flavors.

    197 proslijeđenih tweetova 886 korisnika označava da im se sviđa
    Poništi
  16. proslijedio/la je Tweet
    12. pro 2019.

    For saddle point problems and more general games, gradient descent can cycle and diverge. Using an extrapolation steps adresses this issue and converges to a saddle point (Nash equilibrium).

    6 replies 176 proslijeđenih tweetova 828 korisnika označava da im se sviđa
    Poništi
  17. proslijedio/la je Tweet
    10. pro 2019.

    Found this gem @ ! Using orthogonal property of Legendre polynomials, Legendre Memory Units (LMU) can efficiently handle temporal dependencies spanning 100k timesteps, converge rapidly and use fewer internal state-variables compares to LSTMs.

    11 replies 161 proslijeđeni tweet 681 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    10. pro 2019.

    1st 90+ on GLUE🚨! Baidu’s pre-training language model ERNIE achieves new on (90.1) and becomes the 1st team to elevate the overall score above 90. View more at . Our paper of ERNIE 2.0 has been accepted by .

    23 proslijeđena tweeta 52 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    11. pro 2019.

    Wednesday: Fun day ahead! -9:20 to 10:05 Meet the team at the stand -CAS Score for evaluating gen. models, 10:45 Poster #111 -VQVAE2, 17:00 Poster #140

    23 proslijeđena tweeta 101 korisnik označava da mu se sviđa
    Poništi
  20. proslijedio/la je Tweet
    5. pro 2019.

    We’ve invested in an app for people who text “I’m going to be in LA/SF/NY next week who should I meet?” Currently in TestFlight, growing 30% daily, the round was oversubscribed instantly.

    13 replies 20 proslijeđenih tweetova 355 korisnika označava da im se sviđa
    Poništi

@_hi_o_ 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:

    ·