6 ms·

> It’s important to avoid becoming Gary Marcus and criticize existing technique that work without proposing something else that works even better. Hilarious. H

> It’s important to avoid becoming Gary Marcus and criticize existing technique that work without proposing something else that works even better. Hilarious. Has Gary Marcus actually done anything, in practical terms, like actual code or something, that outperforms the DL approaches he attacks so viciously?

smeeth 6y ago
I keep coming across versions of this comment. Why does it matter? It’s not like the validity or invalidity of Gary’s criticisms would be different if he had done other totally separate ML research.
Rochus 6y ago
Marcus argues convincingly and appears less biased than others who benefit from the current high level of investment in DNN. It seems to be true that no better proposals for solutions have come from his side so far. But I think his criticism per se is valuable, especially his reminder that one cannot simply ignore sixty years of research. EDIT: and of course he is not the only renowned scientist who calls for reflection; here is a quote from an interview with Judea Pearl: "AI is currently split. First, there are those who are intoxicated by the success of machine learning and deep learning and neural nets. They don’t understand what I’m talking about. They want to continue to fit curves. But when you talk to people who have done any work in AI outside statistical learning, they get it immediately. I have read several papers written in the past two months about the limitations of machine learning." (https://www.theatlantic.com/technology/archive/2018/05/machine-learning-is-stuck-on-asking-why/560675/ https://www.theatlantic.com/technology/archive/2018/05/machi...)
- Tenoke 6y ago
  Gary Marcus has been proven wrong on claims he's made and is mostly just nay-saying with adhoc reasoning made up to support it. >appears less biased than others who benefit from the current high level of investment in DNN. Yes, instead he blatantly tries to benefit from the counter-investment in AI skepticism.
  - Rochus 6y ago
    There somehow seems to be an unwritten law of the Python generation like "Thou shalt not criticise Machine Learning". Or is there a better explanation for the emotions that flare up every time someone dampens the exaggerated expectations and reminds us of earlier research in the field of linguistics or AI?
    - Tenoke 6y ago
      There's plenty of good criticism of given products, papers, and approaches. There's also bad criticism though and it's important to distinguish the two.
      - Rochus 6y ago
        The discourse often has little to do with science though.
    - RGamma 6y ago
      Not your problem. Just let 'em be stuck in their local optimum (probably better this way) :P
      - Rochus 6y ago
        > Not your problem Unfortunately I'm not unaffected.
    - YeGoblynQueenne 6y ago
      Let's not make this a generational divide thing, shall we? There's people of all ages who seem to think AI research is like a silly ball game, where one cheers their home team's plays and boos their er, whatsitcalled, the other team's plays, no matter the plays. And who can blame them? AI (read: machine learning) (read: deep learning) research has turned into a huge feeding frenzy. People see all the billions thrown about by Google, Facebook et al. and they go crazy. Maybe they think that if they cheer hard enough and boo hard enough they'll look knowledgeable and "passionate about machine learning" and maybe someone will hire them. Maybe they just want to be on the right side of history, with the winners, not the losers. And when there's so much money to win, there sure are plenty of losers! A while ago someone posted here an article that advised that to become expert in machine learning one should (among other things) "flashcard X papers in major sub-fields" or something along those lines. Pretty revealing of what people are thinking of: Google is hiring machine learning specialists. Shut up Garry Marcus, you'll scare the fish off.
cscurmudgeon 6y ago
I am a DL researcher and practitioner. Any response to criticism that brushes off the criticism based on the source of the criticism is bad and not in the scientific spirit. Progress comes by asking questions and pointing out flaws. You don't need to have an answer to the question before you ask the question. Marcus's criticisms are valid. Most DL researchers have a huge blind spot. It is not good for the field.
YeGoblynQueenne 6y ago
I can't direct a movie but I can tell that Plan 9 from Outer Space is crap (or so crap it's gold, whatever). The kid who pointed out the emperor had no clothes didn't need to be a master tailor. And Gary Marcus doesn't need to propose alternatives to current deep learning practice in order for his criticism of deep learning to be valid. ... and yet he has: neuro-symbolic integration. That's his suggestion. And in fact that's a whole field that's been publishing work for a while now. So he's not just making it up.

smeeth 6y ago

I keep coming across versions of this comment. Why does it matter? It’s not like the validity or invalidity of Gary’s criticisms would be different if he had done other totally separate ML research.

Rochus 6y ago

Marcus argues convincingly and appears less biased than others who benefit from the current high level of investment in DNN. It seems to be true that no better proposals for solutions have come from his side so far. But I think his criticism per se is valuable, especially his reminder that one cannot simply ignore sixty years of research. EDIT: and of course he is not the only renowned scientist who calls for reflection; here is a quote from an interview with Judea Pearl: "AI is currently split. First, there are those who are intoxicated by the success of machine learning and deep learning and neural nets. They don’t understand what I’m talking about. They want to continue to fit curves. But when you talk to people who have done any work in AI outside statistical learning, they get it immediately. I have read several papers written in the past two months about the limitations of machine learning." (https://www.theatlantic.com/technology/archive/2018/05/machine-learning-is-stuck-on-asking-why/560675/ https://www.theatlantic.com/technology/archive/2018/05/machi...)

Tenoke 6y ago
Gary Marcus has been proven wrong on claims he's made and is mostly just nay-saying with adhoc reasoning made up to support it. >appears less biased than others who benefit from the current high level of investment in DNN. Yes, instead he blatantly tries to benefit from the counter-investment in AI skepticism.
- Rochus 6y ago
  There somehow seems to be an unwritten law of the Python generation like "Thou shalt not criticise Machine Learning". Or is there a better explanation for the emotions that flare up every time someone dampens the exaggerated expectations and reminds us of earlier research in the field of linguistics or AI?
  - Tenoke 6y ago
    There's plenty of good criticism of given products, papers, and approaches. There's also bad criticism though and it's important to distinguish the two.
    - Rochus 6y ago
      The discourse often has little to do with science though.
  - RGamma 6y ago
    Not your problem. Just let 'em be stuck in their local optimum (probably better this way) :P
    - Rochus 6y ago
      > Not your problem Unfortunately I'm not unaffected.
  - YeGoblynQueenne 6y ago
    Let's not make this a generational divide thing, shall we? There's people of all ages who seem to think AI research is like a silly ball game, where one cheers their home team's plays and boos their er, whatsitcalled, the other team's plays, no matter the plays. And who can blame them? AI (read: machine learning) (read: deep learning) research has turned into a huge feeding frenzy. People see all the billions thrown about by Google, Facebook et al. and they go crazy. Maybe they think that if they cheer hard enough and boo hard enough they'll look knowledgeable and "passionate about machine learning" and maybe someone will hire them. Maybe they just want to be on the right side of history, with the winners, not the losers. And when there's so much money to win, there sure are plenty of losers! A while ago someone posted here an article that advised that to become expert in machine learning one should (among other things) "flashcard X papers in major sub-fields" or something along those lines. Pretty revealing of what people are thinking of: Google is hiring machine learning specialists. Shut up Garry Marcus, you'll scare the fish off.

cscurmudgeon 6y ago

I am a DL researcher and practitioner. Any response to criticism that brushes off the criticism based on the source of the criticism is bad and not in the scientific spirit. Progress comes by asking questions and pointing out flaws. You don't need to have an answer to the question before you ask the question. Marcus's criticisms are valid. Most DL researchers have a huge blind spot. It is not good for the field.

YeGoblynQueenne 6y ago

I can't direct a movie but I can tell that Plan 9 from Outer Space is crap (or so crap it's gold, whatever). The kid who pointed out the emperor had no clothes didn't need to be a master tailor. And Gary Marcus doesn't need to propose alternatives to current deep learning practice in order for his criticism of deep learning to be valid. ... and yet he has: neuro-symbolic integration. That's his suggestion. And in fact that's a whole field that's been publishing work for a while now. So he's not just making it up.