Mark Bennett

@mpbennett

It's all about you. People-driven results through Oracle Work Life Solutions, Collaboration, & Profiles. Product Strategy and Development

Bay Area, CA
Vrijeme pridruživanja: lipanj 2007.

Tweetovi

Blokirali ste korisnika/cu @mpbennett

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

  1. proslijedio/la je Tweet
    14. sij

    Here's my conversation with Daniel Kahneman, Nobel Prize winning economist and psychologist, author of Thinking Fast and Slow. We talk about two systems of thought in the human mind and their connection to deep learning, AI, and autonomous vehicles.

    Poništi
  2. proslijedio/la je Tweet
    14. sij

    Bernie Sanders is proposing a huge spending increase, without a remotely comparable increase in taxes. Should we care, and if so why? 1/

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    8. stu 2019.

    The Uncanny Vally of ML slides. If you have follow up ideas and questions I'm all ears. Thank you Courtney and team for building such an incredible space for sharing ideas!!!

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

    Call to Action: What if we drew a line and said, we will not let ML systems get worse for the most vulnerable members of our population. Guardrail Metrics fit into our existing systems and easily implemented for drawing that line today.

    Poništi
  5. 8. stu 2019.

    Incentive to decrease contaminates by removal or fine shows results

    Poništi
  6. 8. stu 2019.

    Identifying contamination over time to derive a “CSCORE)

    Poništi
  7. 8. stu 2019.

    Using cameras taking pictures throughout the day

    Poništi
  8. 8. stu 2019.

    Monitoring recycling containers to reduce the costs of bad contents

    Poništi
  9. proslijedio/la je Tweet
    8. stu 2019.
    Prikaži ovu nit
    Poništi
  10. 8. stu 2019.

    Haha. Thank you autocorrect “hair” -> ”gait”

    Prikaži ovu nit
    Poništi
  11. 8. stu 2019.

    Holding a phone, e.g. can alter the analysis when trying to ID by hair

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    8. stu 2019.

    My LSTM model trying to predict the winner of todays giveaway.

    Poništi
  13. 8. stu 2019.

    Identifying humans by their gait “You can tell by the way I use my walk...”

    Poništi
  14. 8. stu 2019.

    Diagnosis moving from expert systems to ML models from Curai

    Poništi
  15. 8. stu 2019.

    In short, selection bias hurts ML and often can’t be seen, we need to understand the collection process better to address that, and use “random” observations to get better insight into selection bias

    Poništi
  16. 8. stu 2019.

    Be “Bayesian” if you don’t know the data collection process

    Poništi
  17. 8. stu 2019.

    2 - what actually gets recorded

    Poništi
  18. 8. stu 2019.

    Selection bias in data collection happens in 2 steps 1 - what happens doesn’t include what could have happened but didn’t i.e. counterfactuals

    Poništi
  19. 8. stu 2019.

    Call to action, as we should always do - protect the vulnerable in society

    Poništi
  20. 8. stu 2019.

    This has a huge impact on our society

    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:

    ·