Tweetovi

Blokirali ste korisnika/cu @heghbalz

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

  1. Prikvačeni tweet
    1. velj

    , Do you know a sponsor for an ML research project that can offer ~50 GPUs for couple of months? Research done at the University, is open source, and we have a convincing proposal. Open to colabs in case needed. Can you share this tweet plz?

    Prikaži ovu nit
    Poništi
  2. prije 1 sat
    Poništi
  3. proslijedio/la je Tweet
    prije 5 sati

    "How to do machine learning efficiently". There's so much to love about this wonderful article.

    Poništi
  4. proslijedio/la je Tweet
    20. stu 2019.

    NEW: Research paper by Florian Henkel, Stefan Balke, Matthias Dorfer and Gerhard Widmer, “Score Following as a Multi-Modal Reinforcement Learning Problem”, using raw score images as input.

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet

    Does anyone know a simple, *fast* way to display a numpy array as pixels, e.g. for watching cellular automata in real time? (I usually use pygame but it has some weird issues on retina displays, so I'm wondering if there's a better option.)

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet

    Wow, I also didn't know about this. I'll try it out next time I'm presenting! "It uses your computer’s microphone to detect your spoken presentation, then transcribes—in real time—what you say as captions on the slides you’re presenting. " (cont)

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    prije 11 sati

    Adding an independent Gaussian variable is the same as doing a heat diffusion on the density.

    Poništi
  8. proslijedio/la je Tweet
    prije 11 sati

    Our white paper from deep fakery workshop is out! It was an amazing experience to collaborate with such a diverse team from legal, social, and computer science backgrounds. The compilation of our observations and our prospective research agenda is below.

    Poništi
  9. proslijedio/la je Tweet
    3. velj

    Alright. Inspired by this tweet, I decided to read all relevant preprints (about 30 in all). Here's what I've found, a thread. 1/15

    Prikaži ovu nit
    Poništi
  10. prije 20 sati
    Poništi
  11. proslijedio/la je Tweet
    2. velj

    Play against my first home made game agent () I built using + + + ! Genuinely amazed that it seems to work 😅 Big thanks my teachers , , , and 👏👏👏

    Poništi
  12. proslijedio/la je Tweet
    prije 21 sat

    speaking about adversarial attacks for invertible generative models at

    Poništi
  13. proslijedio/la je Tweet

    Neural Tangent Kernels (NTK) are a way of understanding neural networks in a certain domain (in the limit of N ->\infty, W0 not ~0). It is beautiful. And somewhat counterintuitive. Here are two nice write-ups aimed at explaining: ,

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

    This is indispensable for quite a few of us! Thanks, for this nifty cheatsheet!

    Poništi
  15. proslijedio/la je Tweet
    3. velj

    what is ? and when a probabilistic model can be deemed tractable? we are brought to divide classes of models into VS ones. A dualism that can be somehow misleading. 1/3

    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    31. sij

    What are the best solns for high level to low level performant creation of RNN/LSTM layers? - Reason: seem unlikely to release new cuDNN RNNs (i.e. LayerNorm LSTM) - The JIT looked promising but JIT LSTM had many problems for me - JAX? TF? Back to CUDA?

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet
    3. velj
    Odgovor korisniku/ci

    Allowing requests for experiments within a one week window favors labs with big compute, and destroys the work-life balance of the people having to run the actual calculations. (You might not be able to start experiments on the _weekend_ the reviews appear, if you have kids.)

    Poništi
  18. proslijedio/la je Tweet
    1. velj

    99 smartphones are transported in a handcart to generate virtual traffic jam in Google Maps. Through this activity, it is possible to turn a green street red which has an impact in the physical world by navigating cars on another route!

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

    My attempt for our book at explaining the difficulty of training deep networks, for folks not familiar with math beyond high school. Any suggestions to make this clearer or simpler (*not* more mathematically rigorous)?

    Poništi
  20. proslijedio/la je Tweet
    1. velj

    Curriculum for Reinforcement Learning "Learning is probably the best superpower we humans have." explores four types of curricula that have been used to help RL models learn to solve complicated tasks.

    Poništi
  21. proslijedio/la je Tweet
    2. velj

    The program chairs answered my inquiry about including requested ablations in the rebuttal. Here’s their official response. Feel free to comment below. Source:

    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:

    ·