Gary Marcus

@GaryMarcus

"If right doesn't matter, we're lost. If the truth doesn't matter, we're lost."

Vrijeme pridruživanja: prosinac 2010.

Tweetovi

Blokirali ste korisnika/cu @GaryMarcus

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

  1. Prikvačeni tweet
    30. pro 2019.

    As this decade comes to a close, and apropos the , I want to repeat one of the first things that I wrote in the last decade, because looking more broadly than we have looked, still strikes me as the best way forward:

    Poništi
  2. proslijedio/la je Tweet

    "You can’t trust this president to do the right thing, not for one minute, not for one election, not for the sake of our country, you just can’t," Rep. Adam Schiff said during his final pitch in the impeachment trial. "He will not change and you know it."

    Poništi
  3. proslijedio/la je Tweet
    prije 2 sata
    Odgovor korisnicima

    the word model gets used in a lot of ways; an ngram model is a model of word usage space but NOT a model of entities in the world; same for Meena, though it is in other ways more sophisticated than n-grams. clear thinking on this point is essential for getting to the next level

    Poništi
  4. proslijedio/la je Tweet
    prije 7 sati

    Does another huge language model prove anything? I spoke to and about Google's new chatbot Meena

    Poništi
  5. proslijedio/la je Tweet
    prije 6 sati

    Agreed. You can't have a meaningful conversation without a conceptual understanding of the world. Let's focus on building that.

    Poništi
  6. proslijedio/la je Tweet
    prije 7 sati
    Odgovor korisnicima

    Meena has exactly the same core issue as ELIZA: it doesn't build a model of what it or the interlocutor has said, and it often contradicts what happened a few turns earlier. Topic without understanding in 1965, topic without understanding in 2020. Here's a sample:

    Poništi
  7. proslijedio/la je Tweet
    prije 17 sati

    ELIZA could do the same, or better, in a tiny fraction of the cost. Over 50 years ago! Yet an example of how is being oversold. is not the solution to all your problems! Proper engineering is: thinks first about requirements and design options!

    Poništi
  8. prije 18 sati
    Poništi
  9. prije 20 sati

    thanks. it also quite relevant to the just-announced Meena...

    Poništi
  10. proslijedio/la je Tweet
    2. velj

    It is amazing that some fathom that GPT-2 just needs more data to reach comprehension and intelligent. well portrays how misguided is this notion.

    Prikaži ovu nit
    Poništi
  11. proslijedio/la je Tweet
    31. sij

    2) the model has 2.5B parameters, 79% score. Adding a sibgle rule to avoid repetition improved about 7% the performance. What do you think twitterspere, can we say that a rule is worth 200M parameters? :) (3/3)

    Prikaži ovu nit
    Poništi
  12. proslijedio/la je Tweet
    2. velj

    whatever Jeep paid Bill Murray for this, it was worth it

    Poništi
  13. proslijedio/la je Tweet

    We’re witnessing the raw, partisan pursuit of power by Senate Republicans, most of whom appear willing to destroy the Constitution if it serves their personal interests. I never thought I’d see such an unpatriotic display in America, let alone the Senate. Never.

    Poništi
  14. proslijedio/la je Tweet
    30. sij

    ROTFL. He really said this? There will be exactly zero Tesla robotaxis by the end of this year and likely same in 2021, 2022 and 2023. I once more reiterate my prediction.

    Poništi
  15. proslijedio/la je Tweet
    30. sij
    Poništi
  16. proslijedio/la je Tweet
    30. sij

    "Prediction does not equal understanding" Great text by NYU professor why we are still far from a real

    Poništi
  17. proslijedio/la je Tweet
    30. sij
    Poništi
  18. 29. sij
    Poništi
  19. proslijedio/la je Tweet

    GPT-2 is fab at generating language -- feed it a sentence, and it'll predict plausible ones that follow, with eerie precision. But does GPT-2 understand the *content* of its prose? No -- because intelligence is not just prediction. By

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    28. sij

    Here's a good reminder that the singularity isn't so near, using my own DeOldify as an example. might like this :)

    Poništi
  21. proslijedio/la je Tweet
    28. sij
    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:

    ·