5 ms·

Reading some Schmidhuber's writings on this issue is a great way to learn the history and development of AI and ANN's that is often missed. At the same time Sch

by nabla9 4y ago

Reading some Schmidhuber's writings on this issue is a great way to learn the history and development of AI and ANN's that is often missed. At the same time Schmidhuber's tone and anal-retentive attitude explains why he is considered so annoying. The field is full of huge egos and overly simplified narratives from a singular point of view. For example, it's true that Hinton was a torchbearer of neural networks for several decades and was off mainstream. That gives him well earned legendary status. But it's also true that other methods worked better until early or mid 2000s and had more rigorous theory behind it (theory behind older methods is still better, nn is still mostly experimenting and wandering). Hinton also had position and funding to do pioneering research and collect students all over the world. I remember reading his product of experts paper as a student in 2002 when I worked with boosting and bagging, so clearly he was not completely out of sight.

fxtentacle 4y ago
I believe it's reasonable for him to request a citation if significant results of his works are quoted. LeCun would probably also be upset if he read about his work in the newspaper, but mis-attributed to someone else. Also, millions in funding money depend on stuff like this. If LeCun can pretend to be the inventor, he gets the grant money. If Schmidhuber gets the citation he seeks, he can use that to apply for new funding. So it's not just prestige, these citation issues have a large monetary impact, too.
- nabla9 4y ago
  [LEC22a] is a position paper. LeCun is expressing his view of how to go forward. It's not a technical paper describing something that works. Ideas he describe have been floating around decades and are considered good ideas, except nobody has made them work, thus the position paper for the future. Schmidhuber points out everything I said above. JEPA is old idea, etc. He also points out that the paper is "tabloid" article. Schmidhuber does not deserve being quoted for general ideas because he wrote paper of few of them.
JohnHaugeland 4y ago
> At the same time Schmidhuber's tone and anal-retentive attitude explains why he is considered so annoying. The people calling him annoying are generally hangers-on trying to display that they're part of the group - people who took a class once 20 years ago and think they need now to come explain. He's a lovely man, and he's 100% correct to be angry that people are getting rich and famous while not citing him. It's an ethical disaster. I really wish HN people wouldn't pile onto him this way. . > The field is full of huge egos and overly simplified narratives from a singular point of view. Stop it. Primacy in publishing is a core topic in science. If these people were at universities, and not Google or Facebook, they'd have been fired for what they're doing. Schmidthuber doesn't actually have a huge ego. He's just saying "guys, stop re-publishing my papers and claiming you came up with it." You should really be deeply disgusted in Hinton right now, not explaining Schmidthuber's ego. Hinton is being a plagiarist *and he knows it*. This isn't an accident and it isn't the first time. Hinton is committing fraud, according to academic doctrine. This is an automatic termination at most universities on Earth, even through tenure. It's time to stop making excuses for these rich men cherry picking other peoples' work for their and their employers' gain. Schmidthuber is, if anything, being remarkably calm and polite, for having been robbed blind by megacorporations of credit and wealth for decades.
- random314 4y ago
  > Hinton is being a plagiarist and he knows it. What are some of the papers he has plagiarized?
  - P-NP 4y ago
    They are listed here: https://people.idsia.ch/~juergen/scientific-integrity-turing-award-deep-learning.html https://people.idsia.ch/~juergen/scientific-integrity-turing...
    - random314 4y ago
      These are some examples of non citations in talks (not papers), not plagiarism. I will discuss a representative example here. > 9. LBH claim ReLUs enabled deep learning to outperform previous methods for object recognition, referring to their GPU-based ImageNet 2012 winner called AlexNet,[GPUCNN4] without mentioning that our earlier groundbreaking deep GPU-based DanNet[GPUCNN1-3,5-8][DAN] did not need ReLUs at all to win 4 earlier object recognition competitions and to achieve superhuman results already in 2011[GPUCNN1-8][R5-6] (see Sec. XIV). If we click and look into the details, Alexnet won imagenet - a general purpose image recognition dataset. Whereas Dannet worked on specific domains- Chinese handwriting recognition, mitosis etc. So Dannet is not comparable in impact to Alexnet at all. ReLUs are in all complex DNNs now - wouldn't have happened if ReLUs are redundant as implied by Schmidhuber. https://people.idsia.ch/~juergen/computer-vision-contests-won-by-gpu-cnns.html https://people.idsia.ch/~juergen/computer-vision-contests-wo...

by nabla9 4y ago

fxtentacle 4y ago

I believe it's reasonable for him to request a citation if significant results of his works are quoted. LeCun would probably also be upset if he read about his work in the newspaper, but mis-attributed to someone else. Also, millions in funding money depend on stuff like this. If LeCun can pretend to be the inventor, he gets the grant money. If Schmidhuber gets the citation he seeks, he can use that to apply for new funding. So it's not just prestige, these citation issues have a large monetary impact, too.

nabla9 4y ago
[LEC22a] is a position paper. LeCun is expressing his view of how to go forward. It's not a technical paper describing something that works. Ideas he describe have been floating around decades and are considered good ideas, except nobody has made them work, thus the position paper for the future. Schmidhuber points out everything I said above. JEPA is old idea, etc. He also points out that the paper is "tabloid" article. Schmidhuber does not deserve being quoted for general ideas because he wrote paper of few of them.

JohnHaugeland 4y ago

> At the same time Schmidhuber's tone and anal-retentive attitude explains why he is considered so annoying. The people calling him annoying are generally hangers-on trying to display that they're part of the group - people who took a class once 20 years ago and think they need now to come explain. He's a lovely man, and he's 100% correct to be angry that people are getting rich and famous while not citing him. It's an ethical disaster. I really wish HN people wouldn't pile onto him this way. . > The field is full of huge egos and overly simplified narratives from a singular point of view. Stop it. Primacy in publishing is a core topic in science. If these people were at universities, and not Google or Facebook, they'd have been fired for what they're doing. Schmidthuber doesn't actually have a huge ego. He's just saying "guys, stop re-publishing my papers and claiming you came up with it." You should really be deeply disgusted in Hinton right now, not explaining Schmidthuber's ego. Hinton is being a plagiarist *and he knows it*. This isn't an accident and it isn't the first time. Hinton is committing fraud, according to academic doctrine. This is an automatic termination at most universities on Earth, even through tenure. It's time to stop making excuses for these rich men cherry picking other peoples' work for their and their employers' gain. Schmidthuber is, if anything, being remarkably calm and polite, for having been robbed blind by megacorporations of credit and wealth for decades.

random314 4y ago
> Hinton is being a plagiarist and he knows it. What are some of the papers he has plagiarized?
- P-NP 4y ago
  They are listed here: https://people.idsia.ch/~juergen/scientific-integrity-turing-award-deep-learning.html https://people.idsia.ch/~juergen/scientific-integrity-turing...
  - random314 4y ago
    These are some examples of non citations in talks (not papers), not plagiarism. I will discuss a representative example here. > 9. LBH claim ReLUs enabled deep learning to outperform previous methods for object recognition, referring to their GPU-based ImageNet 2012 winner called AlexNet,[GPUCNN4] without mentioning that our earlier groundbreaking deep GPU-based DanNet[GPUCNN1-3,5-8][DAN] did not need ReLUs at all to win 4 earlier object recognition competitions and to achieve superhuman results already in 2011[GPUCNN1-8][R5-6] (see Sec. XIV). If we click and look into the details, Alexnet won imagenet - a general purpose image recognition dataset. Whereas Dannet worked on specific domains- Chinese handwriting recognition, mitosis etc. So Dannet is not comparable in impact to Alexnet at all. ReLUs are in all complex DNNs now - wouldn't have happened if ReLUs are redundant as implied by Schmidhuber. https://people.idsia.ch/~juergen/computer-vision-contests-won-by-gpu-cnns.html https://people.idsia.ch/~juergen/computer-vision-contests-wo...