Mike Ross

@MikeNonstop

Chief Scientist at Evid Science — Sculptor — Tennis Fanatic — 🧪 🎨

Los Angeles, CA
Vrijeme pridruživanja: listopad 2011.

Tweetovi

Blokirali ste korisnika/cu @MikeNonstop

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

  1. 6. pro 2019.

    Amazing work from UberAI. They use a discriminator model to steer GPT-2 language generation. So if you can train a classifier to recognize toxic speech (easy), you can now coax your language model to produce coherent non-toxic speech! There are countless use cases for this.

    Poništi
  2. 28. lip 2019.

    Anyone know of papers combining new-school neural NER approaches with old-school gazetteer usage? I'm having trouble finding any contemporary models that discuss gazetteers other than bragging that they don't use them!

    Poništi
  3. proslijedio/la je Tweet
    13. lip 2019.

    We've spent a few evenings last week building an interactive demo called *Write with Transformer* It lets you interact in a very intimate way with GPT-2, call, control, question the model... and I just can't stop playing with it! You can try it at

    Prikaži ovu nit
    Poništi
  4. proslijedio/la je Tweet
    1. svi 2019.

    How do I know so many made-up stories about how the Enigma code was cracked and didn't know until yesterday how interesting the real story is? A volunteer at the National Museum of Computing at Bletchley Park just casually blew my mind with a bit of the story. /1

    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:

    ·