Sivaramakrishnan

@boredsrk

An epsilon in the Doctrine of Chances. Views own. He/him.

Chennai, IN
Vrijeme pridruživanja: siječanj 2018.

Tweetovi

Blokirali ste korisnika/cu @boredsrk

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

  1. proslijedio/la je Tweet

    Statistical mimicry is to language what a plastic toy picturing a bus is to public transportation innovation. It may look right, but it can't perform its function. A horse-drawn cart would be cruder, but could actually move people from A to B. It's cargo-cult AI.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    3. velj

    Added ImageNet validation results for 164 pretrained models on several datasets, incl ImageNet-A, ImageNetV2, and Imagenet-Sketch. No surprise, models with exposure to more data do quite well. Without extra, EfficientNets are holding their own.

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    30. sij

    Did you know that Breiman published both the bagging and random forest papers *after* he retired?!? (This is from our forthcoming book: )

    Poništi
  4. proslijedio/la je Tweet
    29. sij

    A former boss once told me that I should document my lab work because what if I was hit by a bus and they couldn’t continue my project. I never documented anything for that job ever again.

    Poništi
  5. proslijedio/la je Tweet
    28. sij
    Odgovor korisnicima

    The other dimension here is monolingual vs multilingual models. I think monolingual models in low-resource languages currently have an edge e.g. as seen in our MultiFiT paper.

    Poništi
  6. proslijedio/la je Tweet

    Don't assume good health is forever. Don't assume a sharp mind is forever. Don't assume your loved ones are forever. Don't assume you have unlimited time. Do the things you want to do while you have the opportunity, and hug your loved ones close.

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    27. sij

    Amazing: a trail of termites (up) and a trail of ants (down), both protected by a row of their soldiers in a stand-off, without fighting. from

    Poništi
  8. proslijedio/la je Tweet
    25. sij

    When I learn new idea, I must repeatedly learn it from multiple before it *clicks* If you're learning nets - add this to your list It discusses for: - images - audio - databases

    Poništi
  9. proslijedio/la je Tweet
    25. sij

    I think the fact that in France, it's OK to do applied work as long as it is theoretically interesting, and in the US, it's OK to do theory as long as it has practical applications, kind of captures the difference between "Euro PL" and "American PL"

    Poništi
  10. proslijedio/la je Tweet
    23. sij

    Q-learning is difficult to apply when the number of available actions is large. We show that a simple extension based on amortized stochastic search allows Q-learning to scale to high-dimensional discrete, continuous or hybrid action spaces:

    Poništi
  11. proslijedio/la je Tweet
    24. sij

    When your learning rate schedule is perfect.

    Poništi
  12. proslijedio/la je Tweet
    24. sij

    is one of the most important techniques I don't often recommend PhD Thesis' - 's is exceptional. He's a brilliant writer! Check out this taxonomy / table of contents!!! 👇👇👇

    Poništi
  13. proslijedio/la je Tweet
    23. sij

    Hello, world! This is the official account for NumPy, the fundamental package for scientific computing with Python. Follow us for news, info and content related to NumPy!

    Poništi
  14. proslijedio/la je Tweet
    23. sij

    🚀 practicalAI platform launched! - A free tool to 🔍 discover & 📁 organize the top community-curated ML content. I’ll highlight a few of the major features here but be sure to check out the tutorial -

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    21. sij

    Sundar Pichai, CEO of Google, calls for sensible regulation of AI. I agree. "Companies such as ours cannot just build promising new technology and let market forces decide how it will be used. It is equally...

    Poništi
  16. proslijedio/la je Tweet
    22. sij

    I removed a sentence from our 6-page paper being written in LaTeX. So now it is 6.5 pages.

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    20. sij

    I am trilled to see Catalyst usage for Social Good initiatives. This time – smoke prediction to help win the fight against California wildfires crisis. Everything is Open Sourced. Thanks Dmitry Bleklov and for making this happen.

    Poništi
  18. proslijedio/la je Tweet
    30. lip 2018.
    Odgovor korisniku/ci

    Nice list! Our ML lab wrote up a few practical tips for debugging neural networks:

    Poništi
  19. proslijedio/la je Tweet

    most common neural net mistakes: 1) you didn't try to overfit a single batch first. 2) you forgot to toggle train/eval mode for the net. 3) you forgot to .zero_grad() (in pytorch) before .backward(). 4) you passed softmaxed outputs to a loss that expects raw logits. ; others? :)

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    19. sij

    To people tweeting at me saying they can’t get past NYT paywall to read my story. 1. Create an account & you can read some amount of articles w/o paying. OR 2. Subscribe! Wanting everything online to be free is part of what has gotten us into this privacy mess in the first place

    Prikaži ovu nit
    Poništi

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:

    ·