Tweetovi

Blokirali ste korisnika/cu @shivangee_t

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

  1. proslijedio/la je Tweet

    Dismissing machine learning because it can't make sense of what it *hasn't* seen before it quite short-sighted. It is immensely valuable to be able to automatically recognize *what you are able to label* -- especially on hard pattern recognition problems, at super-human accuracy.

    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    10. sij

    Here's the opening lecture I gave on recent developments in deep learning and AI, and hopes for 2020. It's humbling beyond words to have the opportunity to lecture at MIT and to be part of a community that seeks to understand and engineer intelligence:

    Poništi
  3. 24. pro 2019.
    Poništi
  4. proslijedio/la je Tweet
    18. pro 2019.

    It took me a while to recap bc some papers are complex. Key points: - Destructing black box: convergence, generalization, neural tangent kernel - New approaches: Bayesian & uncertainty, graph, convex optimization - Neuroscience & bio-inspired algos - ...

    Poništi
  5. proslijedio/la je Tweet
    16. pro 2019.

    Fun fact: More data can hurt in linear regression. (aside: this is about the 3rd time that studying deep learning has taught me something about linear regression).

    Prikaži ovu nit
    Poništi
  6. proslijedio/la je Tweet
    17. pro 2019.

    graphviz and pydot are fantastic. So I decided to make the most easy-to-use and pythonic library for their functionality that I could. Here is `fastdot`. I hope you like it.

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    16. stu 2019.
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet

    Want to know how adversaries can game explainability techniques? Our latest research - "How can we fool LIME and SHAP? Adversarial Attacks on Explanation Methods" has answers: . Joint work with the awesome team: , Sophie, Emily,

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    28. lis 2019.

    1/ SciTwitter: I'm very excited to share our new Perspective article out in Nature Neuroscience today!

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    3. ruj 2019.

    Introducing Universal Adversarial Triggers Phrases that cause a specific model prediction when concatenated to 𝘢𝘯𝘺 input. Result - GPT-2 turns racist - SQuAD models predict "to kill american people" for 72% of "why" questions - Classifier acc 90%->1%

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    22. kol 2019.

    How long is a "moment"? Moment was a medieval unit of time. 1 day was divided into 24 hours and 1 h was equal to 12 lengths of the period from sunrise to sunset. 1 h was divided into 4 puncta, 10 minuta, or 40 momenta. Considering an average 12 solar hours 1 moment = 90 seconds

    Poništi
  12. 5. kol 2019.
    Poništi
  13. 5. kol 2019.

    Wow, this can make documenting easier and fun!

    Poništi
  14. proslijedio/la je Tweet

    Pakistanis are the friendliest, most hospitable people on the planet. Ever since I landed everyone keeps telling me “apna mulk waapas jao gora” which means “we love you and welcome to our beautiful land.” So happy right now. ❤️

    Poništi
  15. proslijedio/la je Tweet
    26. lip 2019.

    Dyson's law: If AI systems were simple enough to be understandable, they'd be too simple to be intelligent.

    Poništi
  16. proslijedio/la je Tweet
    22. lip 2019.

    And there’s still a gender pay gap.

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet

    Hyperparameter tuning magic

    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    16. lip 2019.

    Boarding a flight and they’re using facial recognition. I had to track down a Delta representative to find out how to opt-out. They did not let travelers know that there was an option to not use it. Again, privacy needs to be the default, we need opt-in systems

    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    12. lip 2019.

    Apparently, GANs are used to create fake profile pictures on LinkedIn for international industrial espionage.

    Poništi
  20. proslijedio/la je Tweet
    10. lip 2019.

    The l^p functional is convex and hence a norm for p>=1. It is sparsity-inducing for p<=1. The l^1 norm is the heart of the lasso, aka basis pursuit.

    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:

    ·