Evgenii Zheltonozhskii

@evgeniyzhe

CS & Physics student. DL researcher. Curious person

github.com/Randl
Vrijeme pridruživanja: travanj 2016.
3 fotografije ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @evgeniyzhe

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

  1. proslijedio/la je Tweet
    14. sij

    OMG! If posted to the today by Ji, Natarajan, Vidick, Wright and Yuen checks out, it's HUGE: a quantum-complexity-theoretic refutation of the Connes embedding conjecture - one of biggest open problems in von Neumann algebras for well over 30 years.

    1 reply 18 proslijeđenih tweetova 60 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  2. proslijedio/la je Tweet
    25. stu 2019.

    I implemented "Momentum Contrast for Unsupervised Visual Representation Learning".

    4 proslijeđena tweeta 19 korisnika označava da im se sviđa
    Poništi
  3. proslijedio/la je Tweet
    9. pro 2019.

    There is no such thing as Artificial General Intelligence because there is no such thing as General Intelligence. Human intelligence is very specialized.

    38 replies 74 proslijeđena tweeta 380 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    12. stu 2019.

    in the future, nobody with h-index less than 15 gets into Stanford

    9 replies 93 proslijeđena tweeta 300 korisnika označava da im se sviđa
    Poništi
  5. 3. lis 2019.

    Writing paper for and want to use natbib textual citations? I've modified the ieee_fullname.bst to support them (thanks to )

    3 korisnika označavaju da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    15. ruj 2019.

    After , the next entry into causal inference is the PRIMER . It was praised already by so many readers, so I won't add, except to note that Wiley is coming up with a clean version next month. In the meantime, the corrected chapters are accessible.

    Hey , I read earlier in the year and it changed my view of everything. What's the next step to getting deeper into Causality and Causal Inference? Which book should we read next?
    23 proslijeđena tweeta 106 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    20. ruj 2019.

    We see more significant improvements from training data distribution search (data splits + oversampling factor ratios) than neural architecture search. The latter is so overrated :)

    45 replies 405 proslijeđenih tweetova 2.031 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    13. ruj 2019.

    New work from our lab out today.
 With & Current compressed sensing (CS) based solutions for MR scan-time acceleration are not practical due to stringent machine constraints that CS doesn’t take into account.

    1 reply 3 proslijeđena tweeta 10 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    15. kol 2019.

    Announcing exciting progress in Bayesian deep learning: the new ATMC sampler achieves first of its kind Bayesian inference results on ImageNet Check out the results and the paper 👇 Heek et al:

    10 replies 89 proslijeđenih tweetova 373 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    14. srp 2019.

    It was "a little bit" hard, but... Finally, Catalyst has full RL algorithmic performance tests! PPO, DQN, DDPG, SAC, TD3 (and distributional improvements) will now be tested on every pull request. That's one small step for framework, one giant leap for reproducible RL research.

    5 replies 31 proslijeđeni tweet 114 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    5. srp 2019.

    As scientific teams grow, our model of credit assignment (1st author, last, or everyone else) becomes increasingly outdated. One impediment is ineffectiveness of author contributions text. Here’s a suggestion for a better way: the contributions table. A thread; feedback welcome.

    634 proslijeđena tweeta 1.761 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. 23. tra 2019.

    Our new paper, "Towards Learning of Filter-Level Heterogeneous Compression of CNNs": tl;dr: we played with differentiable NAS for network compression (quantization and pruning). It turned out to be tricky, unstable and still requires tons of resources.

    1 proslijeđeni tweet 7 korisnika označava da im se sviđa
    Poništi
  13. proslijedio/la je Tweet
    11. tra 2019.

    Just received from SpaceIL communication team what appears to be the last image spacecraft managed to beam to earth before it crashed on the moons surface

    138 replies 1.148 proslijeđenih tweetova 2.629 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    3. tra 2019.

    Very happy to see that the NeurIPS code submission policy explicitly allows code that is not executable "as is." IME as a researcher in industry one of the biggest obstacles to releasing code is decoupling it from non-public supporting infrastructure, so this is great to see.

    15 proslijeđenih tweetova 75 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    4. ožu 2019.

    Model-Based Reinforcement Learning for Atari They show that the simple, iterative method of learning a world model is enough to get ~SOTA data-efficient results on various Atari games. Only 100K interactions between agent and game (2hrs of realtime play)

    8 replies 199 proslijeđenih tweetova 671 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  16. proslijedio/la je Tweet
    3. ožu 2019.

    Are your interested in reproducible RL? Or want a competitive benchmark of current off-policy RL algorithms? check out , catalyst.rl – framework for distributed RL training on top of Various RL algorithms and auxiliary tricks included

    36 proslijeđenih tweetova 109 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. 31. sij 2019.

    Typical data scientist/machine learning engineer: > understands that data science is generally engineering most of programmers can deal with > do want to be overpaid and thus to keep others away from field > keeps saying rare skill he possesses is definitively necessary for DS

    2 korisnika označavaju da im se sviđa
    Poništi
  18. proslijedio/la je Tweet
    9. sij 2019.

    BERT totally fails, btw 😛:

    9 replies 13 proslijeđenih tweetova 81 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. proslijedio/la je Tweet
    6. pro 2018.

    NeurIPS 2018: AI for Prosthetics Challenge – 3rd place solution more info & source code: Want even more info and tricks? NeurIPS workshop RL session is scheduled at Friday at 05:15 PM in Room 518.

    13 proslijeđenih tweetova 35 korisnika označava da im se sviđa
    Poništi
  20. proslijedio/la je Tweet
    3. pro 2018.

    It was a long journey to this commit. Catalyst.RL – distributed training RL framework based on tested on NeurIPS competitions

    50 proslijeđenih tweetova 144 korisnika označavaju da im se sviđa
    Poništi

@evgeniyzhe 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:

    ·