Caroline Lemieux

@cestlemieux

C'est (Caroline) Lemieux. CS PhD Candidate . I like research into making programs more correct. Mostly, I find bugs.

Berkeley, CA
carolemieux.com
Vrijeme pridruživanja: prosinac 2014.

Medijski sadržaj

  1. 25. sij

    Check out the RLCheck preprint! Cool idea to try and use reinforcement learning to tune generators to generate more valid (assumption-satisfying) inputs. With

    1 reply 9 proslijeđenih tweetova 48 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  2. 22. sij

    Finally uploaded video of my talk "Expanding the Reach of Fuzz Testing", which I gave at UMass Amherst! Introduces PerfFuzz, FuzzFactory, FairFuzz, Zest + JQF, RLCheck and Autopandas 😃

    42 proslijeđena tweeta 195 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  3. 27. pro 2019.
    Odgovor korisniku/ci

    Considering that I was coming from this place, Vancouver is *basically* Australia

    1 reply 1 korisnik označava da mu se sviđa
  4. 27. pro 2019.

    Helpful signage in the Vancouver airport.

    16 korisnika označava da im se sviđa
  5. 9. pro 2019.
    Odgovor korisniku/ci

    2 korisnika označavaju da im se sviđa
  6. 26. stu 2019.
    Odgovor korisnicima i sljedećem broju korisnika:

    Twitter IRL

    1 reply 7 korisnika označava da im se sviđa
  7. 26. stu 2019.
    Odgovor korisnicima

    Aghk 😬. Reminds me of this study finding bias in recruitment against people with "foreign-sounding" names... probably quite obvious to anyone who's experienced the bias, but more proof that meritocracy is not reality.

    15 proslijeđenih tweetova 80 korisnika označava da im se sviđa
  8. 17. stu 2019.
    Odgovor korisnicima

    They are quite nice in b&w if I do say so myself

    3 korisnika označavaju da im se sviđa
  9. 1. stu 2019.
    Odgovor korisniku/ci

    This is really cool work! Small comment: FUDGE doesn't do a dynamic analysis of unit tests to get API sequences (as the FuzzGen draft says). It actually does static analysis of library consumers to generate fuzz targets that exercise realistic sequences of API calls.

    1 reply 6 korisnika označava da im se sviđa
  10. 26. lis 2019.

    Headed from Athens to YYZ to go give a talk on research opportunities for PL/SE applied to ML Software Systems. For once, the jetlag will be in my favor for a morning talk on the east coast! >:)

    10 korisnika označava da im se sviđa
  11. 25. lis 2019.

    just finished presenting our paper on Neural-Backed Generators for Program Synthesis, i.e. autopandas, at . Really neat idea on smartly controlling the nondeterminism in generators to do synthesis, if I do say so myself ;)

    3 proslijeđena tweeta 4 korisnika označavaju da im se sviđa
  12. 25. lis 2019.

    And makes it out of a locked bathroom (??) to give the talk on FuzzFactory, which provides a mechanism to build specialized fuzzers by just populating a custom feedback map. FuzzFactory defines waypoints from there!

    1 proslijeđeni tweet 11 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  13. 25. lis 2019.

    Michaël Marcozzi asks whether all this compiler fuzzing really matters.

    4 proslijeđena tweeta 9 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  14. 25. lis 2019.

    Next up, we're not the only ones thinking it's a good idea to combine generator based fuzzing + coverage guided fuzzing! Leo Lampropoulos presents FuzzChick, which combines a pure generative approach for seeding + semantic mutators

    3 proslijeđena tweeta 8 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  15. 25. lis 2019.

    In the test generation session at : "CLOTHO: Directed Test Generation for Weakly Consistent Database Systems". Really cool to see work that's looking at test generation in this space!

    1 proslijeđeni tweet 11 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  16. 24. lis 2019.

    Jan Vitek presenting their reanalysis of "On the Impact of Programming Languages on Code Quality". A great read if you're interested in replication studies!

    1 reply 8 proslijeđenih tweetova 32 korisnika označavaju da im se sviđa
  17. 17. lis 2019.

    We also have a python framework that lets you build your own model-backed generators :)

    3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  18. 17. lis 2019.

    will be presenting our paper at OOPSLA next week!

    1 reply 3 proslijeđena tweeta 7 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  19. 17. lis 2019.

    The idea is pretty cool: use a generator the generate programs in a particular API. To speed up the search, use probabilisitic models bias all the non-deterministic "choice points" in the generator towards generating the correct program.

    1 reply 3 korisnika označavaju da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit
  20. 17. lis 2019.

    When I'm not working on using generators for testing, I work with on using generators for program synthesis. He's giving a tutorial on the technique at RISE camp right now!

    1 reply 1 proslijeđeni tweet 13 korisnika označava da im se sviđa
    Prikaži ovu nit
    Prikaži ovu nit

@cestlemieux još nije poslao/la nijedan tweet s fotografijama ili videozapisima.

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:

    ·