Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Academics, we need to talk

matt-welsh.blogspot.com

Post reply on HN
Print view View original

61–70 of 170 posts

Re: Academics, we need to talk

#61

Post by sanderjd »

kevan wrote: post #43
cushychicken wrote: post #36

Earlier quoted context omitted.

>it's a big deal if a bunch of academics are spending time working on the wrong things At the risk of sounding a little prickly with my comment - isn't it a little presumptuous to write off a certain amount of research as "wrong"? There are plenty of examples of research that didn't have a clear purpose leading to breakthroughs - safety glass, microwave ovens, and Teflon all spring to mind. Also, focusing on efficien…

I understood that as more of a lament that a lot of work is being duplicated because academics aren't aware of prior work in the industry. But then again, how are you supposed to know that someone has a solution unless they tell the world about it in a publication? > It drives me insane to see papers that claim that some problem is "unsolved" when most of the industry players have already solved it, but they didn't h…

There are more ways to "publish" work than in academic publications. I'm not sure to what extent it is unrealistic to expect academics to be keeping up with what is going on at industry conferences or on github or by talking to people at user groups or wherever, but the reality will always be that academic publications are not the only, or the most up-to-date, source of information.

Re: Academics, we need to talk

#62

Post by anonymousDan »

munin wrote: post #29

Today, in academia, it's considered risky to do research in computer vision, machine learning, or speech processing, for example, because it's likely that you will get "out-Googled". Google probably has an entire team of 20 working on what your one graduate student is doing. They'll have petabytes of real data to test against, hundreds of thousands of computers to run their jobs on, and decades of institutional exper…

Such nonsense. I know plenty of people doing great systems research that just doesn't align with the goals of current technology companies. Just look at the proceedings of the top systems conferences and there are plenty of good papers and ideas out there.

Re: Academics, we need to talk

#63

Post by EthanHeilman »

mdwelsh wrote: post #12
forgottenpass wrote: post #9

Now, I don't think all academic research has to be relevant to industry. I'm getting kind of sick of this tactic, where this fake concession is made before going on about how academia isn't designed well enough to deliver to industry. If authors of articles like this didn't care about the efficiency of academia to deliver them research, what's the big deal if there are a bunch of people somewhere running around in ci…

(I'm the author of the original blog post.) Ad hominem attacks aside, I do think it's a big deal if a bunch of academics are spending time working on the wrong things, if for no better reason than maximizing efficiency: Don't forget that most academic research is funded by the government. Since I also happen to help decide which research proposals Google funds, I also care that academics are well-aligned with the pro…

As someone that spent seven years in industry as an engineer and is now three years into a PhD, I lament the failure of industry to adopt old crypto research that has clear immediate application.

I like your post because I think it is the duty of academics to communicate their results for maximum positive effect and to represent real world impact accurately, but I also have to ask what can industry do to expose itself to excellent research which is sitting unused?

Re: Academics, we need to talk

#64

Post by kevan »

sanderjd wrote: post #61
kevan wrote: post #43

Earlier quoted context omitted.

I understood that as more of a lament that a lot of work is being duplicated because academics aren't aware of prior work in the industry. But then again, how are you supposed to know that someone has a solution unless they tell the world about it in a publication? > It drives me insane to see papers that claim that some problem is "unsolved" when most of the industry players have already solved it, but they didn't h…

There are more ways to "publish" work than in academic publications. I'm not sure to what extent it is unrealistic to expect academics to be keeping up with what is going on at industry conferences or on github or by talking to people at user groups or wherever, but the reality will always be that academic publications are not the only, or the most up-to-date, source of information.

Agreed, blog posts and open source projects can be just as helpful as journal-published articles. I didn't intend publications to mean only academic journals.

Re: Academics, we need to talk

#65

Post by wfo »

mdwelsh wrote: post #60
wfo wrote: post #45

If you're doing industry-relevant research and you're in academia, leave. Your work can be supported by corporate profits because it is in essence for corporate profit. Get a job in industry, make more money, and make room for academics who want to do honest-to-god academics and work on theory or fundamental research. Or who want to do research relevant to improving society, not improving profit margins. There aren't…

I don't agree with this at all. The partnership between industry and academia is long-standing and has proven to be extremely valuable -- much of the Internet came about because of it.

At this point it's not a 'partnership' -- industry has co-opted and taken over academia to a startling and troubling extent. Universities are expected to pump out software engineers trained in industry best practice, not thinkers and theorists, and woe betide any school that doesn't toe the line. Research is pushed towards corporate interests as requested in this article here and people just give up. As public research is defunded corporate money has to fill in, which forces academics to make some noises about "applications" or "industry" when they write their papers in hopes of more funding. In any paper, no matter how theoretical, you can make it sound like it has "applications". If someone published the halting problem today, the first half of the abstract would be "as industry deals with more and more complex programs and terabytes of data that are distributed as services in the cloud, we need to understand if there are limits to what we can compute. We present an argument that there are limits, and describe some practical applications".

Many capitulate altogether, trying to do industry work from their academic position. It is these last people who should get out.

The creation of the internet worked wonderfully. It was invented by academics based on government grants when it was basic research, and then refined and turned into something practical and workable and, most importantly, profitable, by industry. Exactly as it's supposed to be.

Re: Academics, we need to talk

#66

Post by jsolson »

swehner wrote: post #39
jsolson wrote: post #25

Earlier quoted context omitted.

That's not really what the article is doing, though... Instead it's saying roughly: much of the work coming out of academia attempts solve problems applicable to industry, but in actuality industry is not actually suffering from the problems solved (either because underlying assumptions are wrong and industry is plagued by a different problem or because the problem has already be adequately solved by existing work).

Some quotes from the article: "My first piece of advice: do a sabbatical or internship in industry." "you have to work on a real product team" "hold yourself to a higher standard" "keep an open mind" Kind of painful to reread it, actually.

All of those quotes are conditioned on the premise that you're doing research which purports to be relevant to industry. In that context I agree with Matt.

> Now, I don't think all academic research has to be relevant to industry. In some sense, the best research (albeit the riskiest and often hardest to fund) is stuff going way beyond where industry is focused today. Many academics kind of kid themselves about how forward-thinking their work is, though. Working on biomolecular computation? That's far out. Working on building a faster version of MapReduce? Not so much. I'd argue most academics work on the latter kind of problem -- and that's fine! -- but don't pretend you're immune from industry relevance just because you're in a university.

This paragraph makes this point abundantly clear: there's plenty of research that is focussed on expanding the field and is discovering problems that will exist once industry catches up. Fantastic. It's risky to try and find the future, but someone has to do it, and I think (based on this and his other writing that I've encountered) Matt would agree that time spent in industry is of dubious utility to those folks[0].

For research that is attempting to tackle extant problems encountered by industry, though, the researcher would be well served by ensuring that the problem they're attempting to solve is actually extant and that the assumptions they're making in trying to solve it don't simply sound reasonable but actually reasonably embody assumptions underlying real (aka industry) deployments. Otherwise they may get a lovely paper which solves a problem that nobody actually has or presents a solution that is impractical because it assumes things which are, in real deployments, incorrect.

[0]: On one hand, it might give them a better feel for in which direction a tractable future lies, but by the same token it might prevent them from exploring potentially fruitful avenues by constraining their thinking.

Re: Academics, we need to talk

#67

Post by adenadel »

zouhair wrote: post #44
mdwelsh wrote: post #12

Earlier quoted context omitted.

(I'm the author of the original blog post.) Ad hominem attacks aside, I do think it's a big deal if a bunch of academics are spending time working on the wrong things, if for no better reason than maximizing efficiency: Don't forget that most academic research is funded by the government. Since I also happen to help decide which research proposals Google funds, I also care that academics are well-aligned with the pro…

I think you are focusing on the wrong problem. The real problem of academics is not what people research, the huge elephant in the room is the "publish or die" dogma. This dogma is one of worst thing happening to research.

I agree that "publish or perish" is not a good model for research. What are alternative models that are potentially viable?

Re: Academics, we need to talk

#68

Post by NotOscarWilde »

_delirium wrote: post #27
jerf wrote: post #14

Earlier quoted context omitted.

"Mostly I'm involved in mobile, wireless, and systems, and in all three of these subfields I see plenty of academic work that tries to solve problems relevant to industry , but often gets it badly wrong." If you want to work on problems irrelevant to industry, go nuts. If you want to work on problems relevant to the industry, it helps to double-check that your problems are relevant to the industry. The desire to work…

Since he doesn't really give specifics, it's hard to tell, but I wouldn't necessarily read most academic papers that mention applications (especially "potential" applications) as having an actual desire to work on short-term industry-relevant problems. Yes, papers often have a vague gesture towards, "breakthroughs in [graph algorithm X] have potential uses in [networking problem]", but this doesn't necessarily indica…

Indeed! The question is: are we ready to drop the requirement of listing applications everywhere?

The department where I am studying is big on graph theory and other theoretical CS concepts, and I always feel a bit uneasy when we write or read about applications (on a grant proposal or in a conference paper, usually) and yet none of the authors really cares about applications. They usually care because "it was an interesting problem that some people studied in the past and we can do it better than them".

Yet, as far as I know, nobody can really write that sentence unless it is such a fundamental problem that its usefulness or importance goes without saying. So we are now in a situation where everyone (theoretical CS, applied theoretical CS, and practical CS) pretends their work have actual applications, and I feel Matt is calling them out on it.

Re: Academics, we need to talk

#69

Post by nickpsecurity »

js8 wrote: post #37
mdwelsh wrote: post #19

Earlier quoted context omitted.

I'm not so worried about duplication by academics -- that does not happen often -- but rather about academic research that's just wrong: makes bad assumptions, uses a flawed methodology, fails to address the general case.

I think I agree with samth - it's not the academia, but the industry who needs to open up. Especially Google has a reputation of being very secretive.

You and samth seem to be missing the point: academic research is often built using methods that will ensure no real-world success while aiming to achieve real-world success. Factoring in real-world constraints, best practices, and existing work-arounds will let academics achieve better results on average. Baseline of practicality goes up.

And again, these are all academic projects that aim at being practical. The point is supported by a number of academics that incorporate real-world information and constraints into their work to produce deliverables that advance state-of-the-art and are useful. Examples that come to mind are: Haskell/Ocaml/Racket languages, CompCert C compiler, Microsoft's SLAM for drivers, SWIFT auto-partitioning for web apps, Sector/Sphere filesystem, the old SASD storage project (principles became EMC), old Beowulf clusters, ABC HW synthesis, RISC-V work, and so on. So many examples of academics keeping their head in the real-world instead of the clouds to making a name for themselves with awesome stuff with immediate and long-lasting benefits. I'm not Matt but I'm guessing he'd rather see more examples like this than, say, a TCP/IP improvement that breaks compatibility with all existing Tier 1-3 stacks and whose goal is to improve overall Web experience. Yes, there are people working on those. ;)

Re: Academics, we need to talk

#70

Post by nickpsecurity »

Fede_V wrote: post #16

I think he is largely correct in identifying the flaws of 'academic research', but he does not spend enough time discussing the whys. Academics are in an insanely competitive environment, where what is rewarded is bringing in grants/high impact publications. There are a very select few academics that are so brilliant and have such sterling reputations they can afford to not play this game (like Matt Welsh's former ad…

This is very important. Writers like Matt might do well to factor in the "academic" constraints much like he encourages them to do for industry. I ran into this in a discussion with Anti on prior art for Rump Kernels. Learned a bit in that discussion but his troubles getting approval stood out:

https://news.ycombinator.com/item?id=10141736

I had heard about BS in academia where only paper output, grants, citations, and so on count. That Anti had to fight to get them to care about his work being implemented speaks wonders about how this works out in practice. Had he only cared about academic success, he could've dropped some light technical details and graphs in the paper then been done with it while the idea collected dust. Many benefited from him fighting the tide on it to produce good papers and an implementation.

I'm not in academic circles but I bet many face the same battle. With the pressure, it might be impossible for them to produce the desired output on their own within their constraints. Or so difficult many give up. Perhaps we should encourage them to have one good line of research they string out over years for quality and ensuring delivery while doing lots of nice papers in between to keep institutions happy. Think that might work?

Post reply on HN