aaron horowitz

@Aaron_Horowitz

Chief Data Scientist , hacker, lover of pancakes. Views are my own.

Vrijeme pridruživanja: svibanj 2011.

Tweetovi

Blokirali ste korisnika/cu @Aaron_Horowitz

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

  1. proslijedio/la je Tweet
    prije 5 sati

    "The diverse forms of disability make it virtually impossible to detect adverse impact with the type of auditing that companies currently use." (Nice work, !)

    Poništi
  2. proslijedio/la je Tweet
    4. velj

    People are very happy to adopt treatments without proper RCTs. But when an RCT suggests a treatment they use does not work, they suddenly become very interested in methodological rigour.

    Poništi
  3. proslijedio/la je Tweet
    2. velj

    When algorithms drive human behavior, feeding the algo fake data can manipulate human behavior.

    Poništi
  4. proslijedio/la je Tweet
    31. sij

    The way we talk about data science and focus so much on methods, we actually incentivize working with *bad* data, rather than spending the time to collect good data and then use easy methods with it

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    31. sij

    Idea: Instead of letting people build more housing, what if we just scream "Housing is a human right" over and over and over again, REALLY loudly?

    Poništi
  6. proslijedio/la je Tweet
    30. sij

    "The story so far: in the beginning packrat was created. This has made a lot of people very angry and was widely regarded as a bad move."

    Poništi
  7. proslijedio/la je Tweet
    30. sij
    Poništi
  8. 30. sij

    - whoever it was suggested that places like the ACLU need more resources for this type of strategic litigation, I agree with the keynote, you'd make a great funding advocate for us!

    Poništi
  9. 30. sij

    We do! You can help us make sure maps across the country are fair, among other things. NBD.

    Poništi
  10. proslijedio/la je Tweet
    30. sij
    Poništi
  11. 30. sij

    - takeaway from explainable ML in Deployment: The reason everyone uses SHAP (wrongly sometimes) is because its **easy to use**. I'd love to see more research in FAT focused on making fairness methods approachable

    Poništi
  12. proslijedio/la je Tweet
    29. sij

    I see this research as "proof of concept" - that light touch interventions can potentially change officer behavior in the field. But more than one experiment is necessary to make policy recommendations.

    Poništi
  13. proslijedio/la je Tweet
    29. sij
    Poništi
  14. proslijedio/la je Tweet
    29. sij

    At when suddenly brings this up, while is also going on... this is why I FOMO this week, twitter friends

    Poništi
  15. proslijedio/la je Tweet
    29. sij

    great work by - overbooking by police (and overcharging by prosecutors) skews justice system in all sorts of ways

    Poništi
  16. proslijedio/la je Tweet
    29. sij

    RStudio Inc is now RStudio PBC. Writing a statement of public benefit and being legally compelled to follow it is a way to reassure the open source community that the right balance will be struck between open source and commercial.

    Poništi
  17. proslijedio/la je Tweet
    29. sij

    I find this idea shared at very true with my experiences, that so many aspects of ML or AI systems impact user experience, even something like a loss function. If we work in data, ML, etc, we are part of the user experience team!

    Prikaži ovu nit
    Poništi
  18. proslijedio/la je Tweet
    28. sij

    ✨ for the folks: is hiring for a summer fellow! do you want to work towards justice in technology, or know someone who does? please reach out and check out the fellowship:

    Prikaži ovu nit
    Poništi
  19. 29. sij

    Bravo to for trying to give the rapidest introduction to pre-trial risk assessments ever provided

    Poništi
  20. proslijedio/la je Tweet
    29. sij

    Today's session 10: The impact of overbooking on a pre-trial risk assessment tool, presented by .

    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:

    ·