Radim Řehůřek

@RadimRehurek

Applied , , . I like to build practical software. PhD in CS. Founder . Creator . Life in SEA 💘 (🇰🇷).

rare-technologies.com
Participa desde outubro de 2013
82 Fotos e vídeos Fotos e vídeos

Tweets

Você bloqueou @RadimRehurek

Tem certeza de que deseja ver estes Tweets? Visualizar os Tweets não desbloqueará @RadimRehurek

  1. Tweet Fixado
    2 de jul

    Scoping and research analysis finished, starting R&D on real data.

    2 respostas 23 retweets 70 curtiram
  2. há 24 horas

    DeepMind tackles another AI barrier: internal environment simulation. Next up in the GAI quest: simulation of self?

    Two new papers describe agents that learn to imagine and use their imaginations to make plans. Read our blog:
    3 retweets 4 curtiram
  3. retweetou
    20 de jul

    Scalnet (keras api in scala) will be more usable next release: great job !

    1 resposta 19 retweets 38 curtiram
  4. retweetou
    19 de jul

    The LSTM Strikes Back: sub 60 perplexities on LM PTB (single model). All you need is a (well tuned) LSTM :)

    4 respostas 112 retweets 333 curtiram
  5. retweetou
    20 de jul

    A call for "engineering corpus", to train domain-specific models on. Do you know such corpus? Please RT.

    1 resposta 14 retweets 17 curtiram
  6. 19 de jul

    A new way to evaluate topic models like LDA: topic intrusion in docs, instead of word intrusion in topics 🎯

    28 retweets 69 curtiram
  7. retweetou
    17 de jul

    Introducing Facets, an open source visualization tool to help understand and analyze datasets

    19 respostas 1.461 retweets 2.322 curtiram
  8. retweetou
    17 de jul
    Em resposta a e

    Thrilled to give a keynote on my fav topic at Warsaw in October: Building practical systems ❤️

    2 respostas 5 retweets 27 curtiram
  9. 15 de jul

    The old curmudgeon has a point though! Just his definition of "life" is too humane and constraining for the voracious AI crowd.

    Rank how well your AI demo went: - Great - Good - Ok - Bad - Very Bad - Miyazaki
    4 curtiram
  10. retweetou
    14 de jul

    Learn & tricks from our developer at this Sunday! 🐍

    5 retweets 18 curtiram
  11. retweetou
    14 de jul

    This is cool. An adversarial attack library

    Foolbox v0.8.0: A Python toolbox to benchmark the robustness of machine learning models
    1 resposta 23 retweets 55 curtiram
  12. 14 de jul

    Already anticipating the startup pitches here 📸 "Instagram Democratized"? "MeUndies for Art"?

    This stuff is crazy! with ML automated image crop, HDR, and masking for amazing pix from panaramas.
    4 retweets 5 curtiram
  13. retweetou
    12 de jul

    I would love to get some feedback from datascientists / kagglers with an extensive experience of feature binning:

    1 resposta 6 retweets 37 curtiram
  14. retweetou
    12 de jul

    Python Implementation of the paper () "Depixelizing Pixel Art":

    11 respostas 200 retweets 458 curtiram
  15. retweetou
    11 de jul

    "TextExt - DBpedia Open Extraction Challenge"

    1 resposta 10 retweets 28 curtiram
  16. retweetou
    9 de jul

    NoScope: 1000x Faster Deep Learning Queries over Video

    1 resposta 11 retweets 34 curtiram
  17. 8 de jul

    Wow, neat extension of the "hashing trick"! Clean and simple idea for Bloom embeddings by HT

    2 respostas 20 retweets 75 curtiram
  18. retweetou
    7 de jul

    "When not to use deep learning" -- Great article!

    4 respostas 224 retweets 440 curtiram
  19. retweetou
    7 de jul

    What are some things you can do to optimize for iteration speed?

    3 respostas 10 retweets 30 curtiram
  20. retweetou
    6 de jul

    TFW a new contributor does an amazing job, fixing a long standing issue 🏅 Congrats and big thanks to !

    1 retweet 20 curtiram
  21. 6 de jul
    Em resposta a

    For the impatient: direct links to the datasets (text similarity, entailment, paraphrasing…)

    1 resposta 16 retweets 33 curtiram

@RadimRehurek ainda não tweetou.

O carregamento parece estar demorando.

O Twitter deve estar sobrecarregado ou passando por algum problema momentâneo. Tente novamente ou acesse o Status do Twitterpara obter mais informações.

    Você também pode gostar

    ·