Carlos Pedrinaci

@cpedrinaci

Applying AI to improve drug development and healthcare. Opinions are my own.

Vrijeme pridruživanja: lipanj 2009.

Tweetovi

Blokirali ste korisnika/cu @cpedrinaci

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

  1. proslijedio/la je Tweet
    17. sij

    Great talk by at : efficient structured debating is very important but currently broken. On the technical side, we need decentralization and interoperability.

    Poništi
  2. 9. sij

    Very useful questionnaire for data science projects/initiatives As usual, great stuff from

    Poništi
  3. proslijedio/la je Tweet
    5. sij

    Consider: millions of years ago our antecedents gave a massive sacrifice of their left hemisphere. We lost a tremendous amount of short term memory and replaced it with Broca’s, Wernicke & the phonological loop. But why? So we can—talk. Thus chimpanzees can do this—we can’t:

    Prikaži ovu nit
    Poništi
  4. 12. pro 2019.

    Interesting analysis of current politics in the UK. No news that economists are at odds with the electorate but strangely people stick to opinions that go against their own judgements during a debate! Defeatism? Inertia? Thoughts? /cc

    Poništi
  5. proslijedio/la je Tweet
    4. pro 2019.

    This looks cool: Allows for library dev in a single notebook with code, tests and documentation in one place

    Poništi
  6. proslijedio/la je Tweet
    5. pro 2019.

    Nice presentation by my colleague Gu Feng tonight at the event at the Gothenburg site. Interested in working with data Science and ai at ”the place where data impacts lives”? Checkout

    Poništi
  7. proslijedio/la je Tweet
    17. stu 2019.

    zzzz  <⌒/ヽ-、___ /<_/____/  ̄ ̄ ̄ ̄ ̄ ̄ ̄ OMG, I didn't turn off the GPU!!    ∧_∧     ( •ω•)   _| ⊃/(___ / └-(____/ oh! It's running on my local machine  <⌒/ヽ-、___ /<_/____/  ̄ ̄ ̄ ̄ ̄ ̄ ̄

    Poništi
  8. proslijedio/la je Tweet
    7. stu 2019.

    . ‘s study tackles the question “How can we harness audience reactions, and pro-actively improve how people engage with emotive video content in a way that develops competencies for peace and prevents genocide?” That’s one hell of a research purpose.

    Poništi
  9. proslijedio/la je Tweet
    10. lis 2019.

    "PyTorch is becoming the deep learning framework of choice at Microsoft" - Big statement from at devcon

    Poništi
  10. proslijedio/la je Tweet
    10. lis 2019.

    Named Tensors are now available in 1.3 ... Love it!

    Poništi
  11. proslijedio/la je Tweet

    Here is a simple experiment for Brexiters: Think of a destination. Get in a taxi. Scream "LEAVE" at the driver. When asked "where to?" respond "JUST GET IT DONE". Let me know where you end up.

    Poništi
  12. proslijedio/la je Tweet

    The idea of a conflict between patriotism and globalism is false and dangerous. 21st century patriotism must have a global dimension, if it is not to become national egoism. The love of one’s homeland cannot become hatred towards one's neighbours.

    Poništi
  13. proslijedio/la je Tweet

    I honestly don’t understand why adults would choose to spend their time mocking and threatening teenagers and children for promoting science, when they could do something good instead. I guess they must simply feel so threatened by us. ->

    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    17. srp 2019.

    Microsoft open-sources scripts and notebooks to pre-train and finetune BERT natural language model with domain-specific texts:

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    7. srp 2019.

    View point invariance is an important inductive bias in how we perceive objects - here tested to the limit by a smart artist.

    Poništi
  16. proslijedio/la je Tweet
    2. srp 2019.

    Most people scan content and make assumptions based on past experiences with similar patterns. A bicycle company struggling with damaged shipments leveraged our propensity to scan by putting a TV image on its box. Damage rates fell 80%.

    Poništi
  17. proslijedio/la je Tweet
    20. lip 2019.

    XLNet outperforms BERT on 20 tasks, often by a large margin, and achieves state-of-the-art results on 18 tasks including question answering, natural language inference, sentiment analysis, and document ranking. Code and comparisons here:

    Poništi
  18. proslijedio/la je Tweet
    11. lip 2019.

    Et voilà! From Jupyter notebooks to standalone applications and dashboards...

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

    *google an issue* * find relevant GitHub Issue* > Ah, 2017! This must've been fixed by now! *scroll through 4 pages* _h4t0n commented 7 days ago_ > any progress on this? GODDAMMIT.

    Poništi
  20. proslijedio/la je Tweet
    10. lip 2019.

    Announcing PyTorch Hub! Want ResNet, ResNext, BERT, GPT, PGAN, Tacotron, DenseNet, MobileNet...? - Pull models with 1 line of code. - Publish your models by sending a pull request. More info here:...

    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:

    ·