[Rant] The Male Only History of Deep Learning

This casting of our field in terms of ideological narrow-sighted group-think is disgusting. Keep Science about ideas!

Links:

[DrDeFrE]

Saying Schmidthuber would approve of an article which attributes GANs to Ian Goodfellow, smh my head...

[RainerZufall-]

When i clicked, I was afraid you would agree with this sentiment... some of these people are just becoming crazy, opening an article on the history of deep learning and focusing on the skin color of the researchers.... That is crazy! He could have named some important research he thought was missing, but only cared about shitting on white people for being white... I'm a leftie, but that is just crazy and I'll never understand such thinking.

But, i don't think "people seem to agree" is quite correct.. it's the people in this guys bubble agreeing, but i don't think this sentiment is that wide-spread in the real world..

[anheuser-busch]

Dont worry, most of us in this field recognize that this is nonsense. Just a weird instance of a bad idea in the real world intersecting with our own. No one cares about the person coming up with the idea. It is the idea itself that matters. Critical thinking is a necessity to be a scientist. This racial/gender importance notion doesnt pass the smell test for anyone who spends their days solving complex problems.

[Notshife]

Could not agree more, the ideas are what matters, doesn't matter who has them

[priyamdey]

Recently revisited this blog post and I see Fei-Fei's picture in the montage which apparently wasn't there before😂 and also "ImageNet" has been added as a part of the chronology

[GarveRagnara]

Thanks for this video. Nowadays, people get celebrated for writing sexist and racist shit like this. It really is disgusting.

[DoubleGlog]

I find it disturbing that in a field as rigorous and evidence based as computer science that there are individuals such as this who entertain this narrative which comes from a narrow, incomplete and inaccurate perspective. That this person has insinuated that somehow there is a pressure within this community to exclude people and be prejudiced on the basis of anything other than merit and accomplishment is disgusting. I work with women, men of all ages and races, and the only thing I will judge them on is the content of their work, not the color of their skin or whatever chromosomes they possess.



I do not for one moment believe that for instance, there is a concerted effort by women to exclude men from the field of nursing and psychology. It happens to be female dominated because of natural preference, and that is okay. The same is true for many STEM subjects. No one should discriminate against anyone except on the basis of their ability to do good work, and I have not seen anything coming close to some kind of discrimination against women or people who have a different skin tone. if anything there has been an overwhelming drive to include women into such fields and they are most welcome. This person is obviously attempting to portray a viewpoint which is heavily skewed and biased at best from one single data point.

[milanpavlovic]

That unmoral person also attack Lecun on twitter for pointing out the solution of problem, also playing on some kind of racial card [in your drama video]. Btw I just found you channel, you are doing nice job man, we're waiting for more advanced topic in AI and more AI memes!

[dylancope]

The article doesn't seem like it's only focused on "algorithmic" changes, it takes about general events and hardware. However, overlooking the importance of datasets and benchmarks is a general problem and I don't think this has anything to do with race/gender.

[clehaxze]

It's science, it's simple. You improve things and make progress, you get recognition.

[TheAnubite]

This is why I love your channel man! You are about keeping it real.

[AnthonyBecker]

I think you're right. This is a silly argument and I appreciate that you are pushing back on it. Perhaps, though, you're not going far enough; we may benefit if you were to take on what these people are *trying* to do, say something about whether you agree or disagree with their motivations, and, if you do agree, what should they do instead of this silliness? Otherwise, perhaps it becomes a cascade of negation... Cheers Yannic.

[mahdinasiri]

MNIST is not on the list!!!! (Triggered)

[dasayan]

Looking back at this video, the tweet was by Timnit Gebru, whos firing from Google made such a big noise in the community

[joshuasmith]

Fei-Fei Li is pretty famous. Not to mention there a ton of famous female research scientists in Deepmind/OpenAI

[PIX_BMS]

If "White and White-adjacent" includes all Caucasians and East Asians, doesn't that describe everyone that's isn't Black?

[josephclements]

Anyone saying imagenet doesn't matter in DL sound be immediately ignored

[alachance]

This problem is infectious and spreading. Identify politics sucks

[XOPOIIIO]

Can you disclose the percent of women viewers of your channel? Thank you.

[mechanicalmonk]

The counter view here about setting a narrative is that the article, by its design and focus, is already presenting a biased view. As you said, ImageNet couldn't have been included because the article focuses only on algorithms, and yet the title isn't "Brief history of deep learning ALGORITHMS", it's simply Brief History of Deep Learning (the author btw states that he doesn't know better and only reported on the history as he found it in mainstream sources which is pretty problematic to begin with).

A history of deep learning simply isn't complete without the contributons of the likes of Fei Fei Li and Daphne Kohler. For us to defend an article that's written with a narrow enough view of the field (i.e. DL == algorithms and that's it) is honestly what's more disgusting.