James

@AwokeKnowing

Software Engineer

California, USA
stackoverflow.com/users/1895611/…
Vrijeme pridruživanja: prosinac 2012.
30 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @AwokeKnowing

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

  1. Prikvačeni tweet
    20. velj 2019.

    We meet new technology with cheers thoughts of utopia or doom and fears As with fire, writing, steel, mathematics with machines, relativity, AI, and genetics the benefits accumulate along with dangers barely understood So let us meet the future with respect and master it for good

    1 reply 2 korisnika označavaju da im se sviđa
    Poništi
  2. proslijedio/la je Tweet
    23. sij

    Google Dataset Search is now officially out of beta. "Dataset Search has indexed almost 25 million of these datasets, giving you a single place to search for datasets & find links to where the data is." Nice work, Natasha Noy and everyone else involved!

    48 replies 2.912 proslijeđenih tweetova 7.629 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  3. 20. sij

    Sideways FTW! Think of system w 64 GPU + model w 64 layers. A GPU handles a layer and passes activations to next GPU AND receives gradient <--back, and calculates backprop based on CURRENT activations, which can be up to 64 frames newer (but like frames in other stage of pipeline

    Given the smoothness of videos, can we learn models more efficiently than with ? We present Sideways - a step towards a high-throughput, approximate backprop that considers the one-way direction of time and pipelines forward and backward passes.
    Poništi
  4. proslijedio/la je Tweet
    27. kol 2019.

    The excellent interactive book "Dive into Deep Learning" has been ported to PyTorch by students at IIT Roorkee The book was authored by , et. al.

    Love this book and great to see customers using it to learn deep learning. Tons of examples, thoughtful discussion, and technical depth. 680 pages and 150 notebooks. Dive into Deep Learning: An Interactive Book with Math, Code, and Discussions
    169 proslijeđenih tweetova 487 korisnika označava da im se sviđa
    Poništi
  5. 26. pro 2019.

    This is so true. So absolutely true.

    Odgovor korisnicima @flcnhvy @Sofiaan @Erdayastronaut
    It’s not that these things can’t be done, they just absorb a lot of software engineering time, which has great opportunity cost. There are very few excellent software engineers & merely increasing team size results in awful bloatware.
    Poništi
  6. 8. pro 2019.

    Just 8 years since DeepLearning arrived at the zeitgeist. Far from just one more tool, DeepLearning set off accelerating development of new & old ideas. Backed by GPU power finally we can generate tractable representations of world around us for further use in many applications.

    Looking for a way to catch up on paper reading before NeurIPS? I've just finished a month-long writing project where I summarized papers on everything from lottery tickets to model-based reinforcement learning, and collected them all here
    Poništi
  7. 26. stu 2019.

    Apparently, vision in the brain has a kind of pretraining, architecture search, phase of development early in fetus development before moving on to regular retinal imaging processing.

    1 korisnik označava da mu se sviđa
    Poništi
  8. 20. stu 2019.

    Flies use a ring of neurons as a compass, with each visual neuron rapidly adapting connection to part of compass to do SLAM. Would love to see a simulation of this.

    1 korisnik označava da mu se sviđa
    Poništi
  9. 3. lis 2019.

    I love it when mathematicians look at various aspects of Deep Learning. Sometimes DL feels like the California gold rush, with a fevered rush to get there first and just start digging. Mathematicians bring a great perspective, like a geologist with instruments bringing sanity.

    Ultra-Wide Deep Nets and the Neural Tangent Kernel (NTK) – Blog | Machine Learning | Carnegie Mellon University
    Poništi
  10. 27. kol 2019.

    Congrats to SpaceX! Looks like we're a go for Multiplanitary Species! We have the tech now. It will only get better from here, as risk and cost go down.

    Dragon Crew Arm. Astronauts will walk this gantry to enter Dragon & depart for the .
    1 korisnik označava da mu se sviđa
    Poništi
  11. 5. kol 2019.

    Best talk in the world on genetic programming. Starts to feel tractable, moving science fiction to practical.

    Tweet je nedostupan.
    Poništi
  12. 27. srp 2019.

    Always remember the baseline.

    2 korisnika označavaju da im se sviđa
    Poništi
  13. 27. lip 2019.

    Great new fast simple python robotics simulator. Super easy to start an RL project.

    0:15
    We are happy to announce PyRep - a toolkit for rapid robot learning research! PyRep is a modification to V-REP that is ~10,000x faster vs the previous remote client approach! Work with & . Report: Code:
    1 korisnik označava da mu se sviđa
    Poništi
  14. 22. lip 2019.

    I would certainly vote for a candidate based on their actions regarding the patent system. Stop patent trolls.

    You can’t patent an abstract idea.  If this bill passes, that will change.
    Poništi
  15. 5. ožu 2019.

    I'm so thankful for the EFF and their years of practical experience. OpenAI’s Recent Announcement: What Went Wrong, and How It Could Be Better via

    1 proslijeđeni tweet 2 korisnika označavaju da im se sviđa
    Poništi
  16. 26. velj 2019.

    I recommend the above slides, along with this great site to software engineers looking to jump into pytorch and AI

    1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. 26. velj 2019.

    Sometimes, diving deep behind the technical curtains makes understanding and using the 'simple' high level concepts much easier. PyTorch under the hood

    1 reply 3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. 9. sij 2019.

    Where do AI scientists go for fun? t-SNE-land!

    2 korisnika označavaju da im se sviđa
    Poništi
  19. 8. sij 2019.

    Should we be concerned the brewing AI patent wars will put the US at a huge disadvantage with respect to China? I'm starting to see a lot of patent applications for things that were released 'publicly' which have numerous github repositories with people implementing them.

    Poništi
  20. 28. pro 2018.

    I think some combination of AlphaZero with Neural Turing Machine would be able to do most of what we want AI to do in the near term, as in complete arbitrary tasks that require following simple instructions. Driving cars, building things at the factory, house cleaning, etc.

    Poništi

@AwokeKnowing 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:

    ·