Richard Dinga

@dinga92

PhD student at VUmc. Brains/psychiatry/evaluation of machine learning models. Deep learning is overrated and discretization of continuous variables is evil

Vrijeme pridruživanja: siječanj 2016.

Tweetovi

Blokirali ste korisnika/cu @dinga92

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @dinga92

  1. proslijedio/la je Tweet
    1. velj

    DataLad: Version control for datasets. > 'DataLad Perpetual decentralized management of digital objects for collaborative open science' by Michael Hanke Prez: Datalad:

    Poništi
  2. proslijedio/la je Tweet
    1. velj

    Now presenting: Michael Hanke DataLad: Perpetual decentralized management of digital objects for collaborative open science Check our live stream:

    Poništi
  3. proslijedio/la je Tweet
    30. sij

    Did you know that Breiman published both the bagging and random forest papers *after* he retired?!? (This is from our forthcoming book: )

    Poništi
  4. 29. sij

    In , instead of a <- 0 b <- 0 c <- 0 you can just do a <- b <- c <- 0 Good for setting up default values of something, or bunch of empty lists/vectors.

    Poništi
  5. 27. sij

    For security reasons, passwords are never sent by email. To set a password, please click this link .... Thank you, a new password was e-mailed to .... 😐

    Poništi
  6. 27. sij

    Give a machine a fish, and you feed it for a day. But try to Max E[fish | state, action] and you will get 1.2B valuation

    Poništi
  7. 25. sij
    Poništi
  8. proslijedio/la je Tweet
    23. sij

    Rosenblatt with the Perceptron computer

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    22. sij

    "If everyone smoked 20 cigarettes a day, then case-control & cohort studies would lead us to conclude that lung cancer is a genetic disease" From this interesting piece on "sick individuals and sick populations": How would mental health fit into this?

    Poništi
  10. proslijedio/la je Tweet
    21. sij

    Tom Nichols () at / today: We reanalysed the same data with several different software packages using parametric and nonparametric analysis, there can be dramatically different outcomes

    Poništi
  11. proslijedio/la je Tweet
    18. sij
    Odgovor korisniku/ci

    Keep throwing this slide up at the end of my job talk...can’t wait to amend that second bullet 😅

    Poništi
  12. proslijedio/la je Tweet
    13. sij
    Prikaži ovu nit
    Poništi
  13. 11. sij

    Terminator is climbing over a pile of human skulls while searching for the last remains of the human resistance. Critics: It's not true AI, because it cannot make a standup routine

    Poništi
  14. proslijedio/la je Tweet

    7 ways to redistrict the entire freakin’ country — from pretty fair to heinously gerrymandered:

    Poništi
  15. 9. sij

    Today I learned that if you write a letter to the editor about an open access paper, you will have to pay $31.50 to read it and another $31.50 to read the response. That's actually a pretty clever business model. Here is a paper for free, wanna know what's wrong with it? Pay us!

    Poništi
  16. 8. sij

    Statistical mistake: computing posterior probabilities given wrong model, arbitrary subjective prior and data, when what you really want is the probability of obtaining test results at least as extreme as the observed results, assuming the null hypothesis. I will hide now.

    Poništi
  17. proslijedio/la je Tweet
    6. sij

    Scaled appropriately, the fire detections are visible without being dramatically so. Burned area shown truthfully is still significant

    Prikaži ovu nit
    Poništi
  18. 6. sij

    If anyone wants to pay me 1M to rerun some algorithms, I might squeeze you in my agenda.

    Poništi
  19. proslijedio/la je Tweet
    6. sij
    Odgovor korisniku/ci

    Virtualised pic of Australia collected from NASA satellites

    Poništi
  20. 3. sij

    Cancer screening is hard even if you are a google AI

    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:

    ·