Sebastian Gehrmann

@sebgehr

NLP+HCI researcher . (1) Shuffle pile of linear algebra. (2) ? (3) natural language. I work on the ❓

Cambridge, MA
Vrijeme pridruživanja: studeni 2013.

Tweetovi

Blokirali ste korisnika/cu @sebgehr

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

  1. Prikvačeni tweet
    6. ožu 2019.

    We developed a tool to evaluate whether a text could have been generated by a language model. This will become increasingly important to prevent abuse.

    Poništi
  2. proslijedio/la je Tweet
    prije 16 sati

    INLG 2020 (International Natural Language Generation conference) will be in Dublin (Ireland), in the week of 7 Sep (this is the week before COLING in Barcelona)

    Poništi
  3. 31. sij

    Forgot to tag as one of the main co-organizers.

    Prikaži ovu nit
    Poništi
  4. 31. sij

    Lots of work went into making this as easy as possible :) Also have a look at the rest of the class on structure and interpretation of deep networks (). Featuring world-class teachers such as and others

    Prikaži ovu nit
    Poništi
  5. 24. sij

    This is based on the code we used for - super useful for any kind of analysis with 🤗.

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

    Here's the complete list of tutorials to be held on Sunday July 5, 2020.

    Prikaži ovu nit
    Poništi
  7. 11. pro 2019.

    Sasha is a fantastic advisor, join him!

    Poništi
  8. 27. stu 2019.

    Barbara has been a fantastic PhD advisor. She doesn't care much about Twitter, but everyone should still read every word she writes.

    Poništi
  9. 30. lis 2019.

    Tagging

    Prikaži ovu nit
    Poništi
  10. 30. lis 2019.

    Best poster of by Amit Moryossef, Ido Dagan, and on planning for Neural data2text generation

    Prikaži ovu nit
    Poništi
  11. 29. lis 2019.

    New paper by salesforce on learning a model-based fact checker for Abstractive Summarization. Definitely a much needed evaluation approach, let's hope that these kinds of metrics will become a new standard. Link:

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

    PyTorch-Struct (v0.3 ). New features: autoregressive models / beam search, sparse-max dp, alignment/dtw, parallel semi-markov, k-max, pretty docs () Fun example: gradients of time-warping crf under different semirings.

    Prikaži ovu nit
    Poništi
  13. 28. lis 2019.
    Poništi
  14. 27. lis 2019.

    Anyone else early in Tokyo for and wants to explore tomorrow? I'm staying near Shinbashi Station. Just shoot me a message and we can meet up!

    Poništi
  15. 26. lis 2019.

    Crowdsource* I'd love an interpretation for why SwiftKey thought I meant to write an adjective.

    Prikaži ovu nit
    Poništi
  16. 26. lis 2019.

    Our ACL tutorial on interpretability got accepted! Please help us crowdsourced potential topics you'd like to see!

    Prikaži ovu nit
    Poništi
  17. proslijedio/la je Tweet

    and present their new framework for collaborative semantic inference at . Better understanding latent variables might allow humans to more effectively incorporate ML tools into their workflows. See more at

    Poništi
  18. proslijedio/la je Tweet
    24. lis 2019.

    Are you interested in becoming a Postdoctoral Fellow? We are now soliciting applications for the 2020-2021 academic year! Please apply through this link: . For more information, visit .

    Poništi
  19. 22. lis 2019.

    Shimabukuro et al.'s H-Matrix at is super interesting. It's an interactive tool to explore typical mistakes that language learners with different native languages make. Available at

    Poništi
  20. 21. lis 2019.

    Crowded room at the workshop at . Check out to see all the work. Everything is an interactive "explainable" aiming to make AI more accessible and intuitive.

    Poništi
  21. 19. lis 2019.

    In other news, I just defended my Ph.D! Thanks to everyone at Harvard, especially my committee members , Barbara Grosz, and . If you want to learn more about Human-AI collaboration for NLG, here are my slides (video coming soon):

    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:

    ·