Phúc Lê

@lkhphuc

Ph.D student at the ML-Labs, Centre for Research Training.

Dublin City, Ireland
Vrijeme pridruživanja: prosinac 2014.
6 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @lkhphuc

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

  1. Prikvačeni tweet
    15. pro 2018.
    Odgovor korisniku/ci

    Hehe I really like the idea that I first heard of from too. So like it that I decided to write an article about it too, even though my writing is not so good.

    18 proslijeđenih tweetova 72 korisnika označavaju da im se sviđa
    Poništi
  2. proslijedio/la je Tweet
    31. sij

    An Opinionated Guide to ML Research: “To make breakthroughs with idea-driven research, you need to develop an exceptionally deep understanding of your subject, and a perspective that diverges from the rest of the community—some can do it, but it’s hard.”

    9 replies 237 proslijeđenih tweetova 851 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  3. 23. sij

    Great thread on Jax and how it’s influenced the development of battle-tested Pytorch.

    Odgovor korisnicima @SingularMattrix @deliprao i sljedećem broju korisnika: 4
    The research hypothesis underlying JAX's design is that there are ever more function transformations to add (pmap, ragged batching w/ mask, fast autoregressive sampling, reversible grad...). So JAX made it easy to add transforms, w/ unified mechanism and very few primitives.
    1 proslijeđeni tweet 4 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. 23. sij

    Very excited with the development for better ML tools.

    Have you ever wondered what will be the ML frameworks of the '20s? In this essay, I examine the directions AI research might take and the requirements they impose, concluding with an overview of what I believe to be two strong candidates: JAX and S4TF.
    1 reply 1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    17. sij

    Planck won the Nobel for discovering quanta of energy. By assuming the energy of light of a certain frequency could only come in discrete steps, he figured out how much light of each frequency you see in a hot furnace. Wild hair. But he was not a revolutionary. (1/n)

    17 replies 148 proslijeđenih tweetova 571 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  6. 14. sij

    I love all features of but I think there are rooms for improvement with the knowledge graph. Or have I been using it incorrectly? Any ideas ?

    3 korisnika označavaju da im se sviđa
    Poništi
  7. 31. pro 2019.

    More discussion.

    Apparently, you need 1-2 publications at top-tier ML conferences to get into a top school for a PhD: I wonder if getting into top schools is still worth it, then? 1/5
    Prikaži ovu nit
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. 31. pro 2019.

    Inspiring! Grateful that do not require half-a-PhD to start a PhD.

    After getting published in ICLR as an Independent Researcher, I have received nearly 100 messages from others who are looking to do the same. So I wrote a blog post on why I decided to do it and my advice to others.
    Prikaži ovu nit
    1 reply 1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  9. 14. pro 2019.

    Cats apparently do pass the mirror test of intelligence.

    5 korisnika označava da im se sviđa
    Poništi
  10. proslijedio/la je Tweet
    6. pro 2019.

    Superb use of scrolling, evoking a gradual descent from sunspeared shallows to inky abyss—and the tenacity with which life permeates the many layers of the planet

    36 proslijeđenih tweetova 92 korisnika označavaju da im se sviđa
    Poništi
  11. 20. stu 2019.

    Great slide of snake oil in the hype.

    Much of what’s being sold as "AI" today is snake oil. It does not and cannot work. In a talk at MIT yesterday, I described why this happening, how we can recognize flawed AI claims, and push back. Here are my annotated slides:
    Prikaži ovu nit
    1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. 18. stu 2019.

    New follow to keep AI hype in check.

    Here we go!
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet
    10. stu 2019.

    😊Self-supervised learning opens up a huge opportunity for better utilizing unlabelled data while learning in a supervised learning manner. My latest post covers many interesting ideas of self-supervised learning tasks on images, videos & control problems:

    11 replies 487 proslijeđenih tweetova 1.797 korisnika označava da im se sviđa
    Poništi
  14. 26. ruj 2019.

    I am using the Internet fast lane by installing 1.1.1.1 app. It's free and protects me from Internet spies!

    Poništi
  15. proslijedio/la je Tweet
    24. kol 2019.

    Protestors in Hong Kong are cutting down facial recognition towers.

    113.576 proslijeđenih tweetova 385.945 korisnika označava da im se sviđa
    Poništi
  16. proslijedio/la je Tweet
    8. kol 2019.

    Temporal Cycle-Consistency Learning (TCC) is a novel self-supervised method for learning representations that are well-suited for fine-grained temporal labeling of video. Learn how it’s done and download the TCC codebase to try it out for yourself!

    233 proslijeđena tweeta 693 korisnika označavaju da im se sviđa
    Poništi
  17. proslijedio/la je Tweet
    30. srp 2019.

    ML dev speed hack #0 - Overfit a single batch - Before doing anything else, verify that your model can memorize the labels for a single batch and quickly bring the loss to zero - This is fast to run, and if the model can't do this, then you know it is broken

    18 proslijeđenih tweetova 100 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    29. srp 2019.

    The recent protests in Hong Kong highlights many potential digital security risks for participants, from electronic payments, smartphone locks, secure chat group management, DDOS attacks from China etc. Article in discusses some of the issues:

    7 proslijeđenih tweetova 27 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    18. srp 2019.

    I'm working on a book on machine learning interviews so I've been spending the last few months talking to companies about their hiring process for ML roles. This thread is a summary of what I've learned. It will be updated as the book progresses. (1/n)

    48 replies 614 proslijeđenih tweetova 2.474 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  20. 8. srp 2019.

    When you visualize your model's embedding with PCA and tSNE.

    0:12
    Va així. És així.
    Poništi
  21. proslijedio/la je Tweet
    6. srp 2019.

    Adamはもう古いのね。cosine annealingがデファクトっぽい / “What’s up with Deep Learning optimizers since Adam?” by

    1 reply 3 proslijeđena tweeta 24 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi

@lkhphuc 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:

    ·