Tweetovi

Blokirali ste korisnika/cu @SmithaMilli

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

  1. proslijedio/la je Tweet
    30. sij

    1️⃣ ‘Value-laden disciplinary shifts in machine learning’ accessibly told “accuracy is not neutral”, illustrated by ImageNet shifting idea of accuracy to favour accuracy in data-intensive, compute-intensive environments:

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

    A thread of some paper presentations that translated between academic disciplines 👇 As an interdisciplinary organisation, we’re excited about work that intersects and communicates across experience, which is crucial but can be under-recognised in academic contexts.

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

    Glad I don't have to keep this a secret anymore - is in Toronto next year! Excited to welcome everyone up to Canada in January ❄️❄️❄️ 🇨🇦🇨🇦🇨🇦

    Poništi
  4. 28. sij

    Flipping the bit isn’t the right approach, but I’m v skeptical that using the pairings from an uninterpretable transport mapping is

    Poništi
  5. proslijedio/la je Tweet
    27. sij

    Twitter Cortex folks and our intern alumna 💫 at

    Poništi
  6. proslijedio/la je Tweet
    25. sij

    Attending FAT*2020? Come to our social at 8:30pm on Tue Jan 28th at the bar of the hotel (conference location). Join us for refreshments, drinks, and lovely company! Allies welcome! 🌈🎉

    Poništi
  7. 27. sij

    Just got to Barcelona for ! Looking forward to seeing everyone :D

    Poništi
  8. 22. sij

    Thanks for having me!! :) really fun to talk to everyone

    Poništi
  9. proslijedio/la je Tweet
    19. sij

    I’ve been finding myself increasingly uninterested in big, sweeping causal questions, eg “how does incarceration affect future employment”. I rarely believe the results of this kind of study. 1/

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    23. srp 2019.
    Odgovor korisnicima i sljedećem broju korisnika:

    Judea, here is work that represents multiples datasets using CI language (DAGS & selection diagrams): & extracts stable ML models. This puts diff methods for identifying stable models into a common framework w/ causal underpinnings.

    Poništi
  11. proslijedio/la je Tweet
    15. sij

    East coast and midwest academic friends! I'll be giving a talk Mon, Nov 9 at UMich & would be happy to give a talk at your uni Nov 10, 11, or 12. DM or email me! More on the book:

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

    Only 3 weeks to the submission deadline! Submit your multidisciplinary work and join us in Addis Ababa! We have our list of speakers up:

    Poništi
  13. proslijedio/la je Tweet
    14. sij

    BREAKING: for the past several years has been relying on undergrads to teach a majority of classes in the Computer Sci (ECCS) department oftentimes for little to NO wages yesterday an arbitrator ruled, at our behest, that UCB OWES the workers more than $5 MILLION

    Prikaži ovu nit
    Poništi
  14. 11. sij
    Poništi
  15. proslijedio/la je Tweet
    9. sij

    Proud to launch the first ever Chinese New Year spot. Proud of the team behind it.

    Poništi
  16. proslijedio/la je Tweet
    9. sij

    What makes big decisions agonizing? What is the computational equivalent of this agony? John McCoy and I reflect on this and other stuff, in a recent symposium on L.A. Paul's () Transformative Experience book.

    Poništi
  17. 3. sij

    I was vegan for 2 years in college and in 2020 I’m doing 90% of the days vegan It’s not 100% b/c I’m skeptical of how nutritionally robust an all-vegan diet is 90% vegan is also relatively easy, it means 36 days to eat whatever, or 3 times a month

    Poništi
  18. 31. pro 2019.

    btw pandora is way better for this b/c they do content-based recs

    Prikaži ovu nit
    Poništi
  19. 31. pro 2019.

    if I listen to e.g. a turkish DJ on spotify, the "related artists" are just other turkish artists (that have nothing in common musically) limits of collaborative filtering! user similarity isn't just caused by music taste, there are confounds like geographic region

    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:

    ·