Chengyu Song

@laosong

Assistant Professor at UCR. Security, System, PL. Gatech & PKU alumnus. What are the important problems in your field?

Riverside, CA
Vrijeme pridruživanja: listopad 2007.
11 fotografija ili videozapisa Fotografije i videozapisi

Tweetovi

Blokirali ste korisnika/cu @laosong

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

  1. proslijedio/la je Tweet
    28. sij

    ☠️ Improperly mixed chemicals cause a shocking amount of injuries in the home and lab ☠️. Thanks to UC Riverside computer scientists, there’s now an app to help prevent reactive chemical accidents.

    1 reply 2 proslijeđena tweeta 9 korisnika označava da im se sviđa
    Poništi
  2. proslijedio/la je Tweet
    28. sij

    BREAKING: We’ve confirmed that the Ring doorbell app on Android covertly shares personally identifiable information on its users with third-party companies, including Facebook.

    247 replies 6.513 proslijeđenih tweetova 6.044 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  3. 10. sij

    I’ve added some new features and a simple afl integration. Also a new name, from Usavich. Check it out. It’s really fast.

    The preliminary code of our fast dfsan-based symbolic executor has been pushed. I probably will create a new repo since we don’t need most code from Angora
    Prikaži ovu nit
    1 proslijeđeni tweet 8 korisnika označava da im se sviđa
    Poništi
  4. 4. sij

    I feel so lucky and proud to be your advisor!

    For promoting rural underprivileged students, I started the Education Foundation in the 2010s w/ friends. We have 40+ in the team, 80% of them were our contestants. We collaborate with 71 schools and cover 20k students. We raised $12K in 2019 for education and humanitarian causes
    Prikaži ovu nit
    1 reply 22 korisnika označavaju da im se sviđa
    Poništi
  5. 27. pro 2019.

    This is the way

    2 korisnika označavaju da im se sviđa
    Poništi
  6. proslijedio/la je Tweet
    13. pro 2019.

    Hoda Naghibijouybari is a PhD candidate at UC Riverside. Her research interests are Architectural Support for Security, GPU Security, Computer Architecture, Heterogeneous Computing. She is interested in faculty and postdoc positions. More here:

    8 proslijeđenih tweetova 13 korisnika označava da im se sviđa
    Poništi
  7. proslijedio/la je Tweet
    11. pro 2019.

    1. We are one of the top computer security labs in the world! 2. We are home of .. 3. We do really interesting and practical Systems & Security research. 4. Also, this: :))

    2 proslijeđena tweeta 13 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  8. proslijedio/la je Tweet
    9. pro 2019.

    A reminder for PhD students and young colleagues, just a reminder that you are not defined by your paper (rejections|accepts). Randomness plays a significant part in most paper/proposal review cycles, and even more so for larger conferences (CHI, security confs, ML conferences).

    9 replies 124 proslijeđena tweeta 531 korisnik označava da mu se sviđa
    Poništi
  9. proslijedio/la je Tweet
    2. pro 2019.

    Thanks again for the invitation and the talks. I can (finally) share the annotated slides with the community: . Hope to see you all at . We are already working on submitting new findings!

    1 reply 11 proslijeđenih tweetova 34 korisnika označavaju da im se sviđa
    Poništi
  10. 1. pro 2019.

    Holiday ends and I have to go back writing papers instead of code 🙁

    1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  11. 1. pro 2019.

    Drawback: compatibility: propagating to libraries is not simple. It really took us quite some efforts to make dfsan work with Google Fuzzer Test suite.

    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  12. 1. pro 2019.

    Quick intro: we use dfsan (source code level taint tracking) to build AST of branch predicate and solve using Z3. Since the AST is “from” LLVM IR, it’s simpler than ones build from binary. Plus it’s source code based taint tracking so I think it will be faster than QSYM.

    The preliminary code of our fast dfsan-based symbolic executor has been pushed. I probably will create a new repo since we don’t need most code from Angora
    Prikaži ovu nit
    1 reply 4 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  13. 30. stu 2019.

    The preliminary code of our fast dfsan-based symbolic executor has been pushed. I probably will create a new repo since we don’t need most code from Angora

    8 proslijeđenih tweetova 39 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  14. 28. stu 2019.

    The preliminary code has been pushed, now it directly solve inside the dfsan_rt, still missing new input generation and fuzzer integration. Coding is fun!!

    2 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  15. 25. stu 2019.

    I was waiting for a 7702p workstation but 3990x? Give me one!

    1 korisnik označava da mu se sviđa
    Poništi
  16. 19. stu 2019.

    Wookhyun is going to release the code and provide a docker. I’m going to use Angora as new base to implement the pure taint based symbolic executor:

    We will release the source code soon
    Prikaži ovu nit
    1 reply 1 proslijeđeni tweet 3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  17. 18. stu 2019.

    Trying on who: 7 min 26 sec, 2963 crashes, 1568 bugs, hit a bug in my code.

    1 korisnik označava da mu se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  18. 18. stu 2019.

    The screenshot shows in 25-30 seconds it can find 56 crashes in uniq from LAVA-M 😊 This is a quick hack based on old implementation based on python z3 binding, I’m going to try native C or rust z3 bindings next

    1 reply
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  19. 18. stu 2019.

    As part of resubmission I just realize we can actually construct symbolic expressions purely based on data-flow analysis ... I'm dumb😓 Results? We can use clean expressions like KLEE while enjoy the speed of QSYM.

    4 proslijeđena tweeta 33 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    17. stu 2019.

    Serious question: Does anyone have suggestions for how to teach "read the error message closely, it tells you exactly where things have gone wrong"?

    10 replies 7 proslijeđenih tweetova 27 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
    Poništi

@laosong još nije objavio nijedan Tweet.

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:

    ·