MIT-IBM Watson AI Lab

@MITIBMLab

A collaborative industrial-academic laboratory focused on advancing fundamental AI research. Affiliated with .

Cambridge, MA
Vrijeme pridruživanja: veljača 2019.

Tweetovi

Blokirali ste korisnika/cu @MITIBMLab

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

  1. proslijedio/la je Tweet
    4. velj

    IBM researchers will share several papers at on , a deep learning design approach could help businesses, non-experts write AI solutions.

    Poništi
  2. 31. sij

    Thanks to all of our Green AI hackers! Five teams won prizes, and all contributed energy-saving ideas that we’re excited to build on. Congrats to , + our other winners.

    Poništi
  3. proslijedio/la je Tweet
    31. sij

    Learn to build an interactive Transformer attention visualization based on and in under 30 minutes! We developed a minimal teaching example for our IAP class, publicly available here:

    Prikaži ovu nit
    Poništi
  4. 30. sij

    Our own speaks to a full house ’s one-week deep learning boot camp. Full syllabus:

    Poništi
  5. proslijedio/la je Tweet
    30. sij

    Meet Natalie Lao, a graduate student in + who wants to make AI accessible for all. via

    Poništi
  6. 29. sij

    Hacking away on the second day of our Green-your- hackathon. Nine teams will present their projects tomorrow. Stay tuned for results!

    Poništi
  7. 28. sij

    A new method for evaluating risk models for specific patients could help doctors avoid ineffective or dangerous treatments. via

    Poništi
  8. 28. sij

    We’re at ⁦⁩ with ⁦’s⁩ ⁦⁦⁩ + ⁦⁩’s Chris Hill learning how to use Satori and run faster, greener AI models.

    Poništi
  9. proslijedio/la je Tweet

    To be truly useful with helping us with the biggest decisions we face today, artificial intelligence needs to reason more like a human. IBM has just taken a big step in that direction.

    Poništi
  10. proslijedio/la je Tweet
    21. sij

    Join us 1/28-1/30 for a three-day tutorial and hackathon! Learn how to use MIT's newest supercomputer, Satori, and how to green up and speed up your . Food and prizes. RSVP:

    Poništi
  11. proslijedio/la je Tweet
    9. sij

    Rapid progress in has fueled a relentless demand for computing power to train more elaborate models on ever-larger datasets. Learn more about how IBM and MIT are responding to this growth in demand.

    Poništi
  12. proslijedio/la je Tweet
    5. sij

    One of my very favorite events in 2019 was the MIT Technology Review in Singapore. It was a pleasure to present my Lab's approach to and discuss it with 's David Cox.

    Poništi
  13. 2. sij

    "We’ve got to improve the computational efficiency of AI so we can do more with it.” + other machine learning experts predict a greater emphasis on AI performance measures beyond accuracy in 2020. via

    Poništi
  14. 20. pro 2019.

    “There’s a ton of text on the internet. Anything to help cut through all that material is extremely useful.” + have a new method to help computers help us find documents we need.

    Poništi
  15. proslijedio/la je Tweet
    17. pro 2019.
    Poništi
  16. proslijedio/la je Tweet
    18. pro 2019.

    I left Harvard almost two years ago, but through the magic of the internet, I'm still teaching there. This year, 163k students are (apparently) taking my course. That's over 1000x the number I taught on campus in my entire time as a professor.

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet

    Why computer vision algorithms need new benchmarks via

    Poništi
  18. proslijedio/la je Tweet
    12. pro 2019.

    [ Workshop Paper Preview] Introducing Memento10k - the largest video memorability dataset to date and MemNet. Cool work by Anelise Newman and colleagues Papers / Schedule at:

    Poništi
  19. proslijedio/la je Tweet
    12. pro 2019.

    I often hear people say that computer vision systems now have "super human" object recognition abilities. They don't. They have have super human abilities *on the ImageNet test set*. Check out "ObjectNet", a new dataset from that shows the gaps:

    Poništi
  20. proslijedio/la je Tweet
    12. pro 2019.

    TODAY: at : Thursday, Dec 12 (THREAD):

    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:

    ·