Ece Kamar

@ecekamar

Artificial intelligence researcher

Vrijeme pridruživanja: kolovoz 2009.

Tweetovi

Blokirali ste korisnika/cu @ecekamar

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

  1. 23. sij

    Do vision-to-language models do as well on reasoning as they do for perception? How can we train for reliability and consistentency on reasoning? Check out

    Poništi
  2. proslijedio/la je Tweet
    23. sij

    At times, we've been impressed with the performance of systems and later discover they've done well for superficial or spurious reasons. With intern we took a closer look at the challenge & directions with deeper DNN reasoning

    Poništi
  3. proslijedio/la je Tweet
    23. sij

    Annoyed that your VQA model can tell whether a banana in an image looks ripe, but can't tell if there is a banana in the image at all? Check out work led by , with wonderful collaborators at (, , , ).

    Poništi
  4. 20. pro 2019.

    We are looking for interns who are excited about exploring the intersection of ML and security on real-world systems. Please apply through the link:

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

    Our job post for internships on Failure Explanation for Machine Learning: . We encourage students working on trustworthy and for to apply and join our team during spring\summer 2020.

    Prikaži ovu nit
    Poništi
  6. 6. stu 2019.
    Prikaži ovu nit
    Poništi
  7. 6. stu 2019.

    I have the pleasure of speaking at the annual Global Challenges 2019 event at this evening. The topic is "Technology, ethics and social change…who cares? " It is open to the public:

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    30. lis 2019.

    On decomposing bias propagation in AI-advised decision making and the impact of equal representation as a bias mitigation strategy in . Andi Peng talking about her residency work in at .

    Prikaži ovu nit
    Poništi
  9. 30. lis 2019.
    Prikaži ovu nit
    Poništi
  10. 30. lis 2019.

    Andi Peng talks about the complex interactions between various sources of bias in hiring decisions @ . The good news is that understanding how bias emerges can help with mitigation. @koriink

    Prikaži ovu nit
    Poništi
  11. 29. lis 2019.

    If we are building AI systems to work with people, why do we think that model accuracy is the only factor that matters? In our recent work published at , we discuss the properties of model error boundary that affect the quality of human decisions on trust.

    Poništi
  12. proslijedio/la je Tweet
    27. lis 2019.

    Now at presents our position paper on a research roadmap for AI Fairness for People w/ Disabilities: (w/ )

    Poništi
  13. 26. ruj 2019.

    Having a great week at the Frontiers of Engineering event organized by the National Academy of Engineering. It is a special pleasure spending time with two other amazing women from

    Poništi
  14. proslijedio/la je Tweet
    17. srp 2019.

    Panelist q and a at session on image captioning . notes "build what is useful, not what is easy... optimize human value, not machine performance"

    Kathy Mccoy, Walter lasecki, ece kamar, and Dan a gurari
    Poništi
  15. proslijedio/la je Tweet
    5. srp 2019.
    Poništi
  16. proslijedio/la je Tweet

    Having a human-in-the-loop means that development is done with a people-first mentality. AI is being crafted carefully by researchers with human outcomes in mind. Dr. shares insights into how AI will evolve in the next century:

    Poništi
  17. 16. lip 2019.

    Today Ada and I are judging wonderful projects from girls who spent their weekend in Seattle hacking for AI for social good. I hope I’ll watch Ada joining this effort herself years from now

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

    Phenomenonal and session by ! Thank you Ece for the unforgettable, inspiring speech - you are a incredible role model and inspiration yourself! 😊🙏

    Poništi
  19. 15. lip 2019.

    I had a great time today talking about responsible AI at hackathon. AI world needs representation of diverse voices to solve hard problems.

    Poništi
  20. 10. svi 2019.

    Congrats to our team winning the best demo award at the DebugML workshop this week.

    Prikaži ovu nit
    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:

    ·