Davide Balzarotti

@balzarot

Professor in Computer Security . Defcon CTF organizer with

Antibes
Vrijeme pridruživanja: siječanj 2011.

Tweetovi

Blokirali ste korisnika/cu @balzarot

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

  1. Prikvačeni tweet
    23. sij

    The Security Circus 2019 update is now online:

    Prikaži ovu nit
    Poništi
  2. 23. sij

    Few highlights: * The number of submissions is back on the exponential curve * It's becoming more and more common for researchers to serve in all PCs in the same year * International collaborations are on the rise. China in particular is growing very fast!

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    7. sij

    Today the Superion fuzzer is not anymore open-source on GitHub () so I created an organization called Fuzzers-Archive that aims to collect unmaintained fuzzers to prevent this situation. Open an issue to ask for additions:

    Prikaži ovu nit
    Poništi
  4. 5. sij

    Articolo assolutamente fuori di testa. La frase migliore (tra tantissime di pregio): "Il tutto naturalmente in Inglese, la vaselina dei popoli" ?!?!

    Poništi
  5. 1. sij

    2020 will be just remembered as the year of bytes/strings encode/decode errors madness

    Poništi
  6. proslijedio/la je Tweet
    31. pro 2019.

    Somehow this just seems to sum up 2019!

    Poništi
  7. proslijedio/la je Tweet
    30. pro 2019.

    Apple: we’re the most secure OS! No bugs, don’t look. Corellium: *provides researches with a proper research environment to find bugs and test their apps* Researchers: *find and report bugs using Corellium* Apple: no no no no nooo NO 😱 *sues Corellium*

    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    24. pro 2019.

    This seems very much the state of ML (and NLP and CV) today. Growth of the field means a collective devaluation of each paper, meaning everyone has to submit more frequently, and get ever closer to the dreaded Least Publishable Unit.

    Poništi
  9. 21. pro 2019.

    The fact that they put the presentations online is great! The fact that they own them and they charge money for it without your consent, not so much...

    Poništi
  10. 21. pro 2019.

    Still got no answer. Were authors asked to sign a document to give ACM the ownership of their presentation? If so, why did they sign ?!?

    Poništi
  11. 19. pro 2019.

    Are video recordings behind a paywall ??

    Poništi
  12. proslijedio/la je Tweet

    For the 30th anniversary of The Cuckoo’s Egg, I interviewed Cliff Stoll and tried to capture the immense, unlikely influence this polymath planetary astronomer has had on the field of cybersecurity:

    Prikaži ovu nit
    Poništi
  13. 11. pro 2019.

    Excellent summary of what's wrong with the current system. But while the problem is clear, the solution is not. After all, this is a case of Goodhart's law: changing the metric to measure success will just bring us to another different (but still likely wrong) state.

    Poništi
  14. proslijedio/la je Tweet
    7. pro 2019.

    Reducing prestige bias in : Single-blind reviewers are significantly more likely than their double-blind counterparts to recommend for acceptance papers from famous authors, top universities, and top companies.

    Poništi
  15. proslijedio/la je Tweet
    4. pro 2019.

    We compiled a long list of Threats to Validity/Relevance in Security Research for our students, building on common sense & our papers on malware experiments and on benchmarking (and inspired by 's excellent blog). Perhaps useful for others?

    Poništi
  16. proslijedio/la je Tweet
    24. stu 2019.

    Grant writing explained | courtesy of for

    Poništi
  17. proslijedio/la je Tweet
    25. stu 2019.
    Poništi
  18. 24. stu 2019.

    Yes, it ticks all the boxes: 1. obtain timestamp, 2. get the idea out there, 3. get feedback + it preserves anonymity. I wish that people would use it more or that arXiv would support a similar model.

    Poništi
  19. 23. stu 2019.

    I think that double blind + anonymous pre-prints (maybe even integrated in hotcrp) could provide a good trade off. Another solution is the current Nature approach: it is up to the authors whether they want to remain anonymous or not (easier but weak solution for reviewers bias)

    Prikaži ovu nit
    Poništi
  20. 23. stu 2019.

    To wrap it up, we have two problems: 1) a slow and random review process that delays research, and 2) biased reviewers who favor well known groups. We can solve one or the other, but solving both in a consistent way is more complex.

    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:

    ·