ImageNet Classification with Deep Convolutional Neural Networks
https://papers.nips.cc/paper/4824-imagenet-classification-wi...
If not evident already, time will tell that this paper brought us to a new era.
Ask HN: What is your favorite CS paper?
221–230 of 265 posts
Re: Ask HN: What is your favorite CS paper?
#222Peter Naur, "Programming as theory building." (1985) “…programming properly should be regarded as an activity by which the programmers form or achieve a certain kind of insight, a theory, of the matters at hand. This suggestion is in contrast to what appears to be a more common notion, that programming should be regarded as a production of a program and certain other texts.” http://pages.cs.wisc.edu/~remzi/Naur.pdf h…
This really is a paper with deep implications. One of them, as I understand it, is that in any significant software project, regardless of the volume and quality of the documentation, or quality of the code base, maintainers not involved in building the original project will not be able to build the "theory" correctly in their minds, and will consequently make changes that are clumsy or detrimental. (I'm summarizing,…
It's at the core of my software life for several years now: http://akkartik.name/about
Re: Ask HN: What is your favorite CS paper?
#223I'll go with an unconventional choice: Michael Kleber's _The Best Card Trick_: http://www.northeastern.edu/seigen/11Magic/Articles/Best%20C...
Re: Ask HN: What is your favorite CS paper?
#224"Reflections on Trusting Trust" by Ken Thompson is one of my favorites. Most papers by Jon Bentley (e.g. A Sample of Brilliance) are also great reads. I'm a frequent contributor to Fermat's Library, which posts an annotated paper (CS, Math and Physics mainly) every week. If you are looking for interesting papers to read, I would strongly recommend checking it out - http://fermatslibrary.com/ - Reflections on Trusting…
Ken Thompson's "Reflections on Trusting Trust" reminds me of this answer to "What is a coder's worst nightmare?": https://www.quora.com/What-is-a-coders-worst-nightmare/answe...
Re: Ask HN: What is your favorite CS paper?
#225Earlier quoted context omitted.
The Ken Thompson Hack is a haunting idea. The intelligence agencies almost certainly would've tried to implement it at some point.
You can actually detect the issue in Trusting Trust: https://www.schneier.com/blog/archives/2006/01/countering_tr... If you have two compilers and one is open source (and you've read the source and happy that it's clean), you can compile that source with both compilers. The output will be different because the two compilers will make different optimizations. However, now you have two binaries of the same compiler and…
You can not. The point is you cannot run the code of a compiler, you have to run the compiled binary. And there's no way to verify if the binary does the same thing as the code when the Ken Thompson Hack is implemented.
Re: Ask HN: What is your favorite CS paper?
#226The Night Watch by James Mickens is always a good read: https://www.usenix.org/system/files/1311_05-08_mickens.pdf
I read The Slow Winter by James Mickens. Reading that brought a revolutionary change in my thinking. Have bookmarked his Harvard profile, just in case he publishes any more articles.
Re: Ask HN: What is your favorite CS paper?
#227One of my all time favorites has been the Jefferey Mogul paper on Receive Livelock: https://pdos.csail.mit.edu/6.828/2008/readings/mogul96usenix...
I read it first as a normal CS paper, but later started seeing it as a commentary on an extremely busy work life.
Right there in the first paragraph: "... receive livelock, in which the system spends all its time processing interrupts, to the exclusion of other tasks..."
Does this remind you of anything?
Re: Ask HN: What is your favorite CS paper?
#228"Hints for Computer System Design" by Butler Lampson
https://www.microsoft.com/en-us/research/wp-content/uploads/...
Re: Ask HN: What is your favorite CS paper?
#229Earlier quoted context omitted.
> I have done this twice, tried different tact twice more, and been downvoted or ignored every time. You should have tried this: "This weird trick by a dad sorts in linear time. Check it out!" Proven to work on so many ad-ridden clickbait websites, so why shouldn't it work on HN? ;-)
"I don't want to live on this planet anymore."
I don't know if your comment is humour in reply to my humour, or if you're actually kinda hurt. So let's err on the safe side.
The short comment "weird trick" that you replied to was just a joke intending to yield a smile to readers, including you. Actually I liked the "speed limit" way you previously used to submit the paper on HN.
That said, I skimmed through the article you mention and found it to look serious and instructive (from my experience getting a Ph.D. in computer science / robotics, yet nothing does not guarantee anything) yet needing to allocate a serious time slot for actual understanding. Many people, even on HN, don't upvote due to complexity, yet it was right to submit it. A number of other insightful comments were written in this thread, thanks for them. Also, your ELI5 explanations are interesting.
My current feeling is like: this sort/discriminator stuff is probably valuable, though it will start usage in demanding situations. It may also eventually be used, without their users even knowing, as a private implementation detail of some data structure in high-level languages. Wait and see.
Back to feelings, this planet has some drawbacks but all in all it's worth it. B612 is too small, we're better here. You can expect good things from HN an similar communities but don't expect too much. Try to refrain from complaining, this feeds the negative part of you and readers as human nature tends to stick bad karma to the ones who complain. Also, when disappointed try to not misattribute causes and favor doubt. Feed the positive part of life.
Re: Ask HN: What is your favorite CS paper?
#230Earlier quoted context omitted.
Are you saying that you can sort strings in O(n log n + D) using a conventional sorting algorithm such as merge sort? If so, I don't understand why D would be an additive factor implying that each string is only involved in a constant number of comparisons. (I wasn't, by the way, only considering strings when discussing variable sized keys -- the beauty of discrimination is that we essentially reduce all sorting prob…
A conventional sorting algorithm such as Multikey Quicksort. http://people.mpi-inf.mpg.de/~sanders/courses/algdat03/salgd...
Sorry, I was being imprecise. By conventional I meant comparison-based sorting functions that are polymorphic in the element type and thus not allowed to examine individual bytes.
Multikey Quicksort indeed looks like a special case of discrimination, exploiting some of the same principles.