peoplewrong

meh my prior on british nationals writing about american decline is not very high...even if he did work for larry summers

the comment I am responding to is talking about computer science

I don't think there are many (any) academic labs in CS with 100 post-docs.

also post docs arent students.

big engineering research efforts like that DO happen in industry though.

I am not really sure how you are defining 'research' but I think your definition might be wildly different from mine. I have a feeling in the pit of my research that what you define as 'practical research' I would NOT define as research in any commonly accepted sense of the word.

> There is a difference between "practical research" where you are developing new ways to solve problems or apply technology and "academic research" where you are developing computer science theory.

this distinction doesn't hold water for me. look at adaboost! but moreover I am very curious what you think 'practical research' looks versus 'academic research'. Like there is a research spectrum in how abstract versus how concrete/empirical computer science research is but 'practical research' is often still pretty complicated/lots of theoretical concerns/ etc etc

> if you want to do practical research, then opportunities are limited if you don't have a PhD

research is hard and it people don't want to hire people for research roles unless they have a track record of doing research. the vast vast majority people with a history of doing 'good' research tend to have Phds. However this is not always the case.

>It doesn't seem like this is the case in other engineering disciplines or even outside of engineering.

I think this is highly unlikely.

who is us? academic computer scientists?

this made me chuckle.

disagreement and people being wildly wildly wrong in good faith [and other people pointing it out!] is an essential part of the scientific process. it is difficult because academics often closely associate themselves with the status of their scientific contributions.

No one really likes killing your heroes but sometimes it must be done.

If psychology claims to be a science it MUST collectively accept this. However we must be sure to deal kindly with the people behind the ideas.

I bet the first version of this was some kind of maximum bipartite matching. like hopcroft-kraft

but on the topic of experts check out this paper by the same research group. This is much more of an 'AI' paper that does include expert domain knowledge and learning

"FutureMatch: Combining Human Value Judgments and Machine Learning to Match in Dynamic Environments" from AAI2015

http://jpdickerson.com/pubs.html

> The kind of system described (though only very vaguely) in the article is a rule-based system, possibly an expert system. These hand-crafted systems were the norm in AI up until the '90s or so and are still in widespread use in various domains (e.g. in finances, or air travel bookings etc).

this is wrong at least for UNOS (united network for organ sharing)-which is the main system in the article-the current computer program used to find matches is an integer linear program. I think it is an interesting question if some kind of rule based system could replicate their ability to find the matches. I tend to think not.

https://pdfs.semanticscholar.org/0837/80a157a7f8ebfa44dce0bc...

page 16 "To our knowledge, there is no solver that would scale to the nationwide steady-state size— including the CMU solver used by UNOS. This solver is based on the work of Abraham et al. (2007), with enhancements and generalizations by Dickerson and Sandholm, and uses integer linear programming (IP) with one decision variable for each cycle no longer than L (in practice, L = 3) and constraints that state that accepted cycles are vertex disjoint. With specialized branch-andprice IP solving techniques, Abraham et al. (2007) were able to solve the (3-cycle, no chains, deterministic) problem at the projected steady-state nationwide scale of 10,000 patients"

edit: wait wait this comment could be out of date. The google scholar for this paper says 2018 but the paper says 2013??

I agree this isn't really "AI" but this project required a lot of algorithm tuning and development....

Here are some good slides.

http://www.cs.cmu.edu/~arielpro/15896s16/slides/896s16-14.pd...

The "credit" should go to the research community that worked on this. which is a mix of economists and computer scientists..........

mechanism design/market design is a huge part of CS that overlaps with economics. Also I am not sure how much the article emphasized this but a big part of the success of these programs is designing 'good' matchings which is computationally tricky and also requires a bunch of domain expertise. actually implementing this and understanding the complexity limitations/what type of ILP (integer linear program)/ fault tolerance etc etc... is a triumph of computer science (also econ!)