Tim Vieira

@xtimv

ML, RL, PL, NLP, bouldering, S&C, parkour, breaking (he/him).

NYC & Baltimore
Joined June 2009

Tweets

You blocked @xtimv

Are you sure you want to view these Tweets? Viewing Tweets won't unblock @xtimv

  1. Pinned Tweet
    5 Nov 2016

    When you say, "This is a reinforcement learning problem," you should say it with the same excitement as "This is NP-hard."

    Undo
  2. Retweeted
    Jan 25

    why i love GridWorld experiments: - finish very quickly - super useful for testing out theoretical results - give great insights into what your algo. is doing - can make them as large as you like (even continuous!) - easy to generate visualizations - they remind me of my PhD days

    Show this thread
    Undo
  3. Retweeted
    Jan 24

    "Backprop is just the chain rule" is what I used to say when I had no idea what I was talking about.

    Undo
  4. Retweeted
    Jan 8

    Fenchel-Rockafellar duality is a powerful tool that more people should be aware of, especially for RL! Straightforward applications of it enable offpolicy evaluation, offpolicy policy gradient/imitation learning, among others

    Undo
  5. Retweeted
    Jan 8

    Can one of you wonderful vision researchers design a model to re-typeset or re-render old math/stats papers w typewritten/hand-drawn notation into a more modern style? Also if you can design a model to reverse the process, that'd be cool but less important. Thanks!

    Undo
  6. Retweeted
    Jan 6

    If you are in the NYC area, the NYAS ML Symposium has an easy abstract submission deadline of Jan 20th. I've found the event a fun way to catch up with local ML folks.

    Undo
  7. Retweeted
    Jan 6

    The FATE group at MSR Montréal has multiple internship positions open for summer 2020. If you are interested in studying the societal implications of artificial intelligence, we are seeking interns who can complement the existing skills on the team.

    Undo
  8. Retweeted
    Jan 2

    got to agree with wikipedia here

    Undo
  9. Jan 2
    Undo
  10. Retweeted
    Jan 1

    As promised, I compiled all my lecture notes on Online Learning in a single PDF. Feedback is welcome! "A Modern Introduction to Online Learning" PS Happy New Year!

    Undo
  11. Retweeted
    Dec 31

    ▓▓▓▓▓▓▓▓▓▓▓▓▓▓▓ 100%

    Undo
  12. Retweeted
    27 Dec 2019

    'Beware of bugs in the above code; I have only proved it correct, not tried it.' -- Donald Knuth

    Undo
  13. Retweeted
    26 Dec 2019
    Undo
  14. Retweeted
    20 Dec 2019

    counterfactuals, what would we do without them

    Show this thread
    Undo
  15. Retweeted

    This is one of the most cursed things I've ever seen: $ python2 >>> import string >>> string.letters 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' >>> help(string) >>> string.letters 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz'

    Show this thread
    Undo
  16. Retweeted
    17 Dec 2019
    Undo
  17. Retweeted
    16 Dec 2019

    Solving linear systems comes with different flavors.

    Undo
  18. 16 Dec 2019
    Undo
  19. Retweeted
    13 Dec 2019

    Want more on measurement, taxonomies of harms, plus examples from fatml + nlp? Check out our tutorial in January with Plus /end self promo

    Show this thread
    Undo
  20. Retweeted
    14 Dec 2019

    Robust regression is obtained by using a l1 loss in place of least squares. Removes outliers. For d=0, computes the median instead of the mean.

    Undo
  21. Retweeted
    12 Dec 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).

    Undo

Loading seems to be taking a while.

Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.

    You may also like

    ·