Florian Tramèr

@florian_tramer

PhD Student in the Applied Crypto Group at Stanford. Interested in Security, Cryptography and Machine Learning

Stanford, CA
Vrijeme pridruživanja: listopad 2019.

Tweetovi

Blokirali ste korisnika/cu @florian_tramer

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

  1. Prikvačeni tweet
    17. sij

    Timing side channels in privacy coins are pervasive! Zcash's recent mobile SDKs are vulnerable to timing attacks similar to ones we (w. ) had found in Zcash & Monero Thanks to for tracking these:

    Poništi
  2. proslijedio/la je Tweet
    3. velj

    We’ve extended the deadline for our ICLR workshop on trustworthy ML to February 12th AoE. The CFP is below. Cc

    Poništi
  3. proslijedio/la je Tweet
    30. sij

    The deadline for our ICLR workshop on trustworthy ML is coming soon on the 31st. Looking forward to your submissions on security, privacy, and other aspects of trustworthy ML! Cc ⁦⁩ ⁦⁩ ⁦

    Poništi
  4. 24. sij

    Best (academic) feeling in the world: realizing that you messed up the conference's LaTeX stylesheet and you actually get an extra page of content with the proper margins

    Poništi
  5. proslijedio/la je Tweet
    23. sij

    Excited to hear from our confirmed speakers Seeta Peña Gangadharan at our upcoming ICLR workshop on trustworthy ML! Submit your work and join us on April 26 in Addis Ababa: Deadline Jan 31 AoE

    Poništi
  6. proslijedio/la je Tweet
    21. sij

    Submissions to our ICLR workshop on trustworthy ML are now accepted! Details on the call for papers and way to submit can be found on the website: Deadline: Jan 31 AoE

    Poništi
  7. 13. sij

    Always a cool event! I'll be talking about the side-channel attacks we recently discovered in Zcash and Monero

    Poništi
  8. proslijedio/la je Tweet
    7. sij

    The call for papers is up for our ICLR workshop on trustworthy machine learning: Submission deadline: Jan 31, 2020 AoE Looking forward to your papers! cc

    Poništi
  9. proslijedio/la je Tweet
    18. pro 2019.

    I have multiple open positions for graduate students, visiting students, and postdocs to work on security & privacy of machine learning. Reach out if you are interested or know someone that I should contact!

    Prikaži ovu nit
    Poništi
  10. proslijedio/la je Tweet
    17. pro 2019.

    Nice comprehensive look at the state of federated learning research, including an overview of advances in the last three years (since in 2016), as well as a nice overview of open problems in the area. 105 pages, 485 references!

    Poništi
  11. proslijedio/la je Tweet
    12. pro 2019.

    New paper on "Advances and Open Problems in Federated Learning", together with 58 co-authors from 25 institutions:

    Prikaži ovu nit
    Poništi
  12. 12. pro 2019.

    I'll present our paper on "Adversarial Training and Robustness for Multiple Perturbations" at 10:35AM in West Exhibition Hall A Swing by poster #87 right after if you'd like to learn more or to chat about ML security more broadly paper:

    Poništi
  13. proslijedio/la je Tweet

    Right now AdBlock Plus uses a simple list of perceptual hashes, but I expect ad blocking will evolve to use machine learning on visual features. A recent paper argues that this method is brittle, so it'll be fascinating to see the arms race in practice. 🍿

    Prikaži ovu nit
    Poništi
  14. proslijedio/la je Tweet

    Today's talk is given by Dmitry Kogan. He'll discuss private information retrieval with sublinear online time. Now, this is not the problem of preserving data privacy. Rather, it is the problem of querying data without anybody knowing what was queried.

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    8. pro 2019.
    Odgovor korisnicima i sljedećem broju korisnika:

    By the way, the biggest remaining side-channel-leakage in Zcash is the network layer (inherited from Bitcoin).

    Poništi
  16. 5. pro 2019.

    Finally got to write a hyped² paper on blockchain and ML!

    Poništi
  17. proslijedio/la je Tweet
    3. pro 2019.

    I am co-organizing a workshop with Nicholas Carlini and at ICLR 2020. The theme will be around trustworthy ML, more information is available on our (preliminary) website:

    Prikaži ovu nit
    Poništi
  18. 25. stu 2019.

    The recent progress in text generation is of course impressive, but it feels like we're getting dangerously close to having to replace the "infinite monkey theorem" with the "infinite refreshing of theorem"

    Prikaži ovu nit
    Poništi
  19. 25. stu 2019.

    Q: GPT-2, were your answers cherry-picked? A: Of course, there is only so much that you can find out from a 5-year-old source: (without any cherry-picking this time)

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    22. stu 2019.

    Much clicked * v1 * Eth2 client update * how reduced call volume * Convexity Protocol * remote sidechannel attacks on Monero/ZCash * taxonomy of LAOs * tokenizing offers on

    Tweet je nedostupan.
    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:

    ·