Gabriel Brat

@bratogram

Trauma surgeon, surgical informatics, and co-founder. Tweets are my own.

Vrijeme pridruživanja: siječanj 2018.

Tweetovi

Blokirali ste korisnika/cu @bratogram

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

  1. 29. sij

    Great kick-off lecture by course director . Excited to teach Health IT innovation course at with for the sixth year!

    Poništi
  2. proslijedio/la je Tweet

    Hey friends, is anyone aware of a typology of data visualization users? Has anyone studied how much variation there is in how effectively different audiences use visualizations? Either for general visualization problems or specific applications. Thank you!

    Poništi
  3. proslijedio/la je Tweet
    12. srp 2019.

    Practical problem: you have a dataset and want to do causal inference. How to report the validity region, i.e. where the new policy should be used? Joint w/ , Dennis Wei, , Tian Gao, , (part of the ):

    Poništi
  4. 6. srp 2019.

    On the positive side, the ratio of death to coverage for trauma was not the worst.

    Poništi
  5. proslijedio/la je Tweet
    3. lip 2019.

    Is there [medical] value in AI prediction? Come hear the panel [June 18th] incl Christine Tsien Silvers @gbratMD discuss this at annual Patient-driven Precision Medicine

    Poništi
  6. proslijedio/la je Tweet
    1. lip 2019.
    Poništi
  7. proslijedio/la je Tweet
    26. ruj 2018.

    Evolution of Hokusai's "Great Wave". 1. When he was 33 (1792). 2. When he was 44 (1803). 3. When he was 46 (1805). 4. When he was 72 (1831).

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    28. lip 2018.

    The National Trauma Registry Repository. "The NTRR was developed by trauma researchers for trauma researchers." Read about this exciting development at

    Poništi
  9. proslijedio/la je Tweet
    17. svi 2018.

    Opioid Prescribing Education in Surgical Residencies: A Program Director Survey, most residents prescribe opioid, few receive training, Proud to train our residents on

    Poništi
  10. proslijedio/la je Tweet
    15. svi 2018.

    Typical oncology office visit note for one of my patients: 3880 words, 21490 characters total (w/o spaces) The useful part of this note, not repeating information that is better-displayed elsewhere in : 86 words, 489 characters ⇒ 98% of this note is garbage

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    17. tra 2018.

    Adversarial examples have become a major area of AI research, but their greatest practical ramifications could turn out to be in medicine. We make the economic+technical case, providing context for both CS and medical readers: w/ +

    Poništi
  12. proslijedio/la je Tweet
    5. tra 2018.

    Our preprint that creates embeddings for over 100k medical concepts using data from 60 million patients, 1.7 million journal articles and 20 million notes is up: Pretrained embeddings: Interactive explorer

    Prikaži ovu nit
    Poništi
  13. proslijedio/la je Tweet

    this... may be the greatest story ever told

    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet
    4. tra 2018.

    Nice example of correlation vs causation. In this case, correlation of dog death with use of is high. But causation appears to be a third factor: United are the only carrier who will accept high-risk dog breeds (and they warn customers when they do so).

    Poništi
  15. proslijedio/la je Tweet
    18. ožu 2018.

    And just the beginning of the most dramatic changes medicine and healthcare have ever seen. Up to us to be sure it is for the better. And moves medicine from paternalism to partnership; from averages to individuals. With people fully in charge of their data and decisions.

    Poništi
  16. proslijedio/la je Tweet
    14. ožu 2018.

    Check out our new paper on "Spatial Graph Convolutions for Drug Discovery." Converts a 3D macro molecular structure into a graph structure that it feeds into a graph convolutional deep network. Matches state-of-art with end-to-end learning.

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet

    we now a sufficiently large dataset to show conclusively (p < 1E-80) that thoughts and prayers do not work

    Poništi
  18. 18. velj 2018.

    Great article: Choice of fluid type: physiological concepts and perioperative

    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:

    ·