Judea Pearl thinks that DL is nothing more than curve fitting. Is he a greivence scholar?
In the humanities criticism plays a similar role.
A researcher criticizing slavery or modern social structures isn't completely now free of any standards.
The journal they tricked has an impact factor of 1.1, so im not impressed.
Using 'greivence scholars' is stupid, and contributes to the STEM superiority complex.
What everyone means when they refer to "grievance scholars" is the same group of people that were popularly called S.J.W. (social justice warrior) only a year ago, except with less popular baggage and a more academic tone (which might not last once popular personalities get a hold of it). I don't know what the technical requirements for something to be a real label are, but I can confidently say that it is a real group of people. As reluctant as we all are to admit it, I think the "Jordan Peterson" reactionary movement has made the jump from YouTube personalities to academia, and although they might not want to be associated with their origin story, all of the language and concepts are coming along.
>Judea Pearl thinks that DL is nothing more than curve fitting. Is he a greivence scholar?
This is a red herring, because although many off-mainstream scholars have literal grievances, their primary subject of study is not the fact that they do.
I agree essentially, its a meme created to debase the humanities.
And I'm not too worried about the legitimacy of this viewpoint from within the academic community.
What is dangerous is that people who are only marginally invested form a dangerous and disingenuous opinion of large fields of research on the basis of "SJWs are annoying".
> On Pearl
Obviously a counterexample within the context of STEM. What this guy hates about the humanities is that it challenges the status quo.
The point I'm trying to make with this relates to communities and the perception of criticism.
Foulcaut would say all the humanities are capable of is to criticize modern institutions.
> I'm just talking about the modern context.
When we say "greivence studies",
This is a phrase this guy came up with, to obviously target a subset of the humanities with a left leaning adgenda (i.e. womens studies, african american studies, and whatever gay studies is called), and target a low quality journal to confirm his preheld belifes.
It's not a useful idea to try to understand discourse or solve any problem within the humanities, but an disrespectful attempt to mar whole departments.
The choice of the label Grievence scholar is deliberate. If to critically understand anything is impossible, all you're doing is complaining.
> Marring whole fields of study
This is the intended response. Not to expose flaws in the paper to publication pipeline. So people could point and laugh.
It doesn't help that STEM people love to do this anyway.
Until you bring up research fraud in their respective field, and how dumb did that made them all look.
In bio this happens all the time, most often at the top of the research hierarchy.
Associating "grievance studies" with the humanities might be a little too broad, after all the ultimate villain of modern critical thought (Colonial-era affluent Englishmen, probably), loved and generously funded the humanities of their day. To cast the claim in the strongest light I would say that it simply refers to the minority of people that, under the flag of helping, have made horrible accusations towards innocent people (a brief review of the literature will turn a few of these up, after all there's nothing stopping them from being published). In the most pessimistic light, "grievance studies" would just be, as you say, a conservative action word belonging right next to all of the similar liberal action words. The fact of the matter is that it is probably being used right now in both ways, by different camps.
Edit to respond to edit in the parent comment:
>but an disrespectful attempt to mar whole departments.
The claim that "whole departments" deserve to be marred is exactly what he's saying, and a lot of people agree. In this comment thread, I saw someone call the group of people (which out of respect for your argument against using the term I will not name) a "gangrenous foot." There is no word-mincing going on when someone calls whole departments gangrenous feet.
>> Grievance scholars isn't a real label
> I agree essentially, its a meme created to debase the humanities
I think you're agreeing with the quote of your own comment.
>I agree essentially, its a meme created to debase the humanities.
Stop trying to associate humanists who care about literature and art, and social scientists who care about truth, with political activists who have academic cover.
They didn't target one journal, they targeted multiple. 7 papers were published and 5 others were in the process of being iterated on, all taken seriously. They cited numerous existing papers to jump off of and build their ridiculous premises. Such as the idea that white students should be chained up in class to understand privilege or that obesity is great. These papers were lauded as exciting contributions and one was singled out with an award.
This work is nearly indistinguishable from real papers in the field, and was constructed to justify ridiculous conclusions, working backwards to find precedent in the fields in question. That's what makes them qualify as Grievance Studies. There are no real hypotheses being tested, it's just the academic laundering of ideology.
Do you seriously think someone could spend a year just learning the lingo of a real STEM discipline and get similarly farcical work published? With no real results required? This isn't about real sociology, linguistics, literature or psychology.
The fact that these disciplines are actively trying to break into STEM by claiming a necessity for feminist geography or astronomy really says enough. They're not happy to just sit in their own departments, doing "research" and producing an inbred body of work, they want to compel the hard sciences to take them seriously.
Remember the C+equality parody that went around a few years ago and which ticked off enough gender activists to get it banned from github? Well, they were just lampooning a real thesis someone had written, about how feminism could enrich logic and coding, because masculinity focused too much on rigid categories. This empress already has no clothes.
As a matter of fact, yes.
Google scigen, mathgen - you will find the stories of how computer-generated nonsensical papers were accepted by conferences and low quality journals.
For example: http://thatsmathematics.com/blog/archives/185
It seems to me (as was actually pointed out by the hoaxing authors themselves) that either you can take what they did seriously, or not, but not both.
"If the members of your committee of inquiry object to the very idea of satire as a form of creative expression, they should come out honestly and say so. But to pretend that this is a matter of publishing false data is so obviously ridiculous that one cannot help suspecting an ulterior motive."
In my opinion, this equally applies to allegations of unethical experimentation.
I think it would be possible to publish a paper in one of the large machine learning venues with completely made up results and techniques, that nonetheless looked very impressive. One would have to take a bit of care to make the paper look legitimate, in other words, copy the style and conventions of a machine learning paper- but it's not that hard to do. After all, a great deal of machine learning research is, well, not quite made up, but for example there are important details left out of papers, experiments are poorly described, data is not available, results are interpreted in fanciful manner, etc etc.
Of course, such a "hoax" would be immediately condemned by the whole field because results would have to be fabricated. Then again, that is what Boghossian is accused of.
But DL nets can be very fragile solutions, that only curve fit to the training data.
You dont want to recognize the same picture of the puppy, you want to recognize all puppies.
Others who disagree are trying to integrate logical approaches with DL into novel archetectures.
I see it as more of pendulum moving back and forth.
But Pearl's work is in Causal Statistics and is a hard core Bayesian.
https://www.quantamagazine.org/to-build-truly-intelligent-ma...
It's not like his complaints aren't being addressed in the field, they are just not easy problems (transfer learning, world models)
In a neural analogy, you don't haven't built a brain, you have a reflex.
Machine learning is rapidly losing its ability to innovate as a field and the reason is the extreme focus on deep learning for classification- and secondarily the fact that the billions of corporate funding now attract researchers from the sciences who have no background in computer science or AI and haven't got a clue what the fuss is about that people like Perl are making.
Here it is from Hinton himself:
One big challenge the community faces is that if you want to get a paper published in machine learning now it's got to have a table in it, with all these different data sets across the top, and all these different methods along the side, and your method has to look like the best one. If it doesn’t look like that, it’s hard to get published. I don't think that's encouraging people to think about radically new ideas.
Now if you send in a paper that has a radically new idea, there's no chance in hell it will get accepted, because it's going to get some junior reviewer who doesn't understand it. Or it’s going to get a senior reviewer who's trying to review too many papers and doesn't understand it first time round and assumes it must be nonsense. Anything that makes the brain hurt is not going to get accepted. And I think that's really bad.
https://www.wired.com/story/googles-ai-guru-computers-think-...
You'd probably be best served by reading Pearl himself, his recent Book of Why is good, this much earlier paper[0] kind of lays out a few things too as for his reasoning that causal inference needs different tools. There's also some stuff in the blogosphere about an earlier dispute you might find fun, i.e. the methods of using causal inference tools. Example[1].
[0] https://ftp.cs.ucla.edu/pub/stat_ser/r284-reprint.pdf
[1] https://statmodeling.stat.columbia.edu/2009/07/05/disputes_a...
Deep learning? It is curve fitting (ie pattern recognition), is there supposed to be something wrong with that?
What are the high impact journals of their fields? My understanding is that Hypatia is a leading journal.