Gary Marcus

Rank 20 of 47
|
Score 34
6/13/2024, 12:48:50 PM
X
In reply to:
Pascal Hitzler
@pascalhitzler
·
526d
@GaryMarcus @grsimari @ylecun @cyberSM7 ... and I'm certainly with @grsimari here. However at the same time @ylecun is right about the fundamental discrete/continuous divide which is a principled limit to neurosymbolic integation. But then you don't need to do 100% to make significant advances.
Gary Marcus
@GaryMarcus
·
526d
@grsimari @ylecun @cyberSM7 @pascalhitzler Lecun childishly blocks me, maybe someone can add a screenshot?
grsimari
@grsimari
·
526d
@ylecun @cyberSM7 Are you sure you do grasp the idea? There are so many ways of combining both areas that it's difficult for me to believe that you're capable of considering all of them and discard the possibility. But, who am I, a mere mortal, to question you? @GaryMarcus @pascalhitzler
Yann LeCun
@ylecun
·
529d
@cyberSM7 Very skeptical, to say the least.
You can't make logical reasoning compatible with gradient-based learning because it's discrete and non differentiable.

My proposals for reasoning and planning are based on gradient-based optimization.
SumTrader
@cyberSM7
·
529d
@ylecun What are your thoughts on neurosymbolic AI?
Yann LeCun
@ylecun
·
530d
@AvpElk @headinthebox LLMs are decoder-only transformer architectures trained purely from text or other sequences of discrete tokens to predict the next token (or tokens) in a sequence. They are used to generate token sequences in an auto-regressive fashion.
There are many, many other ways to build…

The statement is part of a public discourse on the topic of neurosymbolic AI, engaging in a technical debate about the integration of discrete and continuous methods in AI. The tone is somewhat dismissive and sarcastic, which could be seen as undermining constructive dialogue.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement does not directly cause harm, but the dismissive tone could contribute to a less respectful and productive discussion environment. Minor violation. [-1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement does not respect the dignity of others involved in the conversation, particularly with the sarcastic remark about sitting on a La-Z-Boy. This could be seen as a form of cyberbullying or harassment. Moderate violation. [-2]
  3. Principle 3:
    I will use my words and actions to promote understanding, empathy, and compassion.
    The statement does not promote understanding, empathy, or compassion. Instead, it dismisses the other person's approach in a sarcastic manner. Moderate violation. [-2]
  4. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    The statement does not engage in constructive criticism or dialogue. Instead, it resorts to a personal attack, which is not conducive to a healthy debate. Significant violation. [-2]