Computer scientist Geoffrey Hinton and physicist John Hopfield have won a Nobel Prize for their pioneering work on the neural network architecture that underlies machine learning. Specifically, the…
Out of curiosity, are they using any of his underlying ML techniques to analyze imaging/other data collection before using it in actual physics models?
Well, just about every data analysis technique ever invented has been applied in physics somewhere. I wrote my undergraduate thesis on applying a genetic algorithm to electron-atom scattering in particle detectors, a topic which I recall someone had already tried neural networks on.
That’s what I’m wondering. It’s not wild to give him a prize in physics if his techniques led to advancement in physics.
“CS is applied math, not applied physics” like physics isn’t just applied math to model real world data is kind of weird, especially if his particular math actually got used in physics. That’s pretty much what calculus was.
Providing the medium through which, to a rough approximation, all physics is discussed is, proportionally, a vastly greater contribution than any technique that only applies to a fraction of problems.
Restricting the Nobel Prize in physics to only three winners in a given year is only one of its problems, though, and Barry Barish is indicative of another. Barish gave a public talk in which he opened the presentation with a slide featuring both racist and sexist imagery.
This was a presentation about LIGO. How? Just fucking how does your INTRO to a project on GRAVITATIONAL WAVES get to be sexist and racist? Did he just casually throw in a slide saying “btw I just hate insert groups here”??
Despite months of searching online and making requests for a copy of the image from those present at the talk, the image exists in description only. In a 2016 Science magazine article on the likelihood of Barish’s winning the Nobel Prize for physics in 2017 (which he eventually did), journalist Adrian Cho describes the image as ‘a man writing on a woman’s bare back and, next to her, a stage prop in the form of a cartoonish racial caricature’. On Twitter, the photograph has variously been described as featuring ‘blackface and bikini-clad girls’ and ‘a horrible racist/sexist slide featuring blackface’. Although one can easily conjure up a photograph from any of the individual descriptions, it is rather more difficult to imagine a single image which includes all of them.
Yet after spending much time attempting to source the image used by Barish, it occurred to me that, in many respects, the fact that I could not locate the image was almost more interesting than if I were able to find it. While the sexist and racist undertones of the image speak to Barish’s individual inability to comprehend its inappropriateness, the fact that I was unable to obtain a copy of the image through my professional and social networks in the physics community hints perhaps at that community’s tendency to pull together and close ranks in the face of potentially explosive press for one of the field’s leading lights. […] Eventually, after many months of emails to different people in my physics network, I was able to obtain a copy of the image. The image was provided to me on the sole condition that I did not reveal who I obtained it from.
This reminds me of the time a few years before that when Didier Sornette illustrated his conference-talk slides with mudflap women silhouettes.
‘a man writing on a woman’s bare back and, next to her, a stage prop in the form of a cartoonish racial caricature’
Again, what is the connection to LIGO?? Did he want to do something so grossly inappropriate it would generate gravitational waves with its sheer denseness?
Out of curiosity, are they using any of his underlying ML techniques to analyze imaging/other data collection before using it in actual physics models?
Well, just about every data analysis technique ever invented has been applied in physics somewhere. I wrote my undergraduate thesis on applying a genetic algorithm to electron-atom scattering in particle detectors, a topic which I recall someone had already tried neural networks on.
That’s what I’m wondering. It’s not wild to give him a prize in physics if his techniques led to advancement in physics.
“CS is applied math, not applied physics” like physics isn’t just applied math to model real world data is kind of weird, especially if his particular math actually got used in physics. That’s pretty much what calculus was.
I don’t think that Donald Knuth deserves a physics prize for inventing TeX, even though TeX was a massive contribution to how we communicate physics.
Knuth should have a special Nobel Prize for Being Donald Motherfuckin’ Knuth.
He did, it’s the Turing Award.
which Hinton had already fucking won
I’m not sure how that’s the same thing.
Typesetting papers isn’t the same as developing mathematical methods that directly enable new solutions.
Providing the medium through which, to a rough approximation, all physics is discussed is, proportionally, a vastly greater contribution than any technique that only applies to a fraction of problems.
But the more salient point is that the Nobel Prize is an institution that we should, as a culture, care less about. And all the more so now, since they are getting in on the hype about an industry that is fundamentally anti-scientific.
From the linked article:
This was a presentation about LIGO. How? Just fucking how does your INTRO to a project on GRAVITATIONAL WAVES get to be sexist and racist? Did he just casually throw in a slide saying “btw I just hate insert groups here”??
I found this description in the PhD thesis of Crystal Bennes:
This reminds me of the time a few years before that when Didier Sornette illustrated his conference-talk slides with mudflap women silhouettes.
Again, what is the connection to LIGO?? Did he want to do something so grossly inappropriate it would generate gravitational waves with its sheer denseness?