The papers published on the Google File System (GFS) and Map Reduce are still some of my all time favorite papers. It gives really good inside into how GFS/Map Reduce was built, but explains it in a very straight forward way. We actually implemented an in memory version of GFS/MapReduce in my graduate operating systems class. It remains as one of my favorite projects I've ever done.
In the vain of generally appreciated papers, I really like "Whitesides' Group: Writing a Paper" by George Whitesides[0][1]. It gives a strategy for collaborating research based on using a paper as a living document. It seems like a lot of work, but it saves untold days in the long run. This is the first paper I give anyone I mentor.
There have been derivative works on giving presentations, that I also particularly like: Editorial: Effective Presentations—A Must. [2]
[0] In case you don't know of him: he is the most cited living chemist, or something to this effect
Why: RPC and its ilk make a lousy model for mobile data, since mobile devices are only occasionally connected, not permanently. Similarly, in the face network and server failures, servers can be modeled as occasionally connected as well. The "replication" mindset is far more productive when dealing with those issues. The linked paper gives a broad overview of a great number of approaches to replication, and is a great way to get the lay of the land.
This paper kickstarted the concept of information theory, and was hugely influential on many fields of research. Signal-to-noise ratio, the bit, information entropy, etc. are all theories and concepts presented by Shannon.
Wow you beat me to it! This paper affected my college major, my career in my twenties, and my current career. Such an important paper, although, I must admit, it is not easy to decipher at first.
This was the first academic paper I ever read, way back when. For a brief, wonderful time, I thought all academic papers were of a similar caliber, and I was in awe (and not a little bit intimidated).
> For a brief, wonderful time, I thought all academic papers were of a similar caliber
I had the exact same thought in college when I started reading a handful of well-written papers like this myself. This really set a standard that for 60+ years now academia hasn't always maintained.
+1! I find the whole notion of "typical set" to be absolutely amazing.
I would like to share with you a few pages from the intro to my thesis which cover Shannon's channel coding theorem. There are some nice TiKZ illustrations. http://minireference.com/static/excerpts/Shannon_channel_cod... (it's not super detailed, but the definitions of all the moving parts are given)
This is really cool. I'm working in systems research, but I'm fascinated by mathematical research in CS. Did you continue with similar work after your PhD, If I may ask?
I've since switched my research focus to machine learning (look up latent Dirichlet allocation, very cool stuff). I find a lot of parallels between the two fields: prob. theory, matrices, uncertainty, ...
I'm still following quantum information theory research, but more as a spectator from the sidelines. However, a couple of weeks ago I had to come back to quantum info. theory to "defend" my academic reputation. Our colleagues from TIFR found a bug in one of our papers (http://arxiv.org/abs/1111.3645v3) so my coauthor and I had to fix it. It was kind of cool to see I hadn't "lost my quantum skills" after two years of running a business. I guess, once you go quantum you never go back? :)
Very cool! Shannon's paper was, to be honest, a little out of my comfort zone when I first read it (this isn't really my field academically), and I got a little lost on the first read of this, but someone below posted your full thesis which I will definitely read when I have the time.
Stephen Jay Gould and Richard C. Lewontin, 1979 "The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme"
Some biologists may cringe (especially the Gould haters), but I don't think I've ever been so engrossed by any other scholarly paper. It is a joy to read. Very approachable for non-biologists. The papers critique of sloppy "just so" reasoning, could easily be extended to Data Scientists/Engineers/Entrepreneurs. Highly recommend!
I am cringing. Gould was a classic case of projection, having been guilty of everything he accused his opponents of: misreading one's opponents, proneness to ideological bias, and experimental technique so sloppy that deliberate fraud starts to look like the simpler explanation. (http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjo...)
There are other bones to pick with Gould, but these are the ones that make him impossible to read as an interested layperson without personally verifying every sentence.
William P. Thurston 1994, "On proof and progress in mathematics"
Gives a good amount of insight into how academia works for mathematics, and gives a good contrast with how CS works. Don't be scared by the abstract, it's a completely non-technical paper. The academic/research culture can be more important than the results.
For something completely different, Amelia Rauser, "The Butcher-Kissing Dutchess of Devonshire: Between Caricature and Allegory in 1784." Eighteenth-Century Studies 36 (Fall 2002): 23-47.
"The Letter S" by Donald E. Knuth. An entire paper on the typographical design of the letter S and variants based on type size and other attributes. An elegant paper on a single letter:
I can't tell if you're serious but...I can get it for free if I just wait to be in the office tomorrow (we subscribe to springer). If I remember that is.
However, many people here are not going to pay the money or have access through their employer, and will miss out. Free Knuth!
Since someone already posted "Spandrels", I'll go with "How Not to be a Bioinformatician" by Manuel Corpas, Segun Fatumo, and Reinhard Schneider[0]. It's a humorous takedown of very common problems in Bioinformatics. I think the point comes across better when you say "if you do X, you are doing poorly" versus "don't do X if you want to do well". Plus, it's a little cathartic.
Came here to say this. It really is awesome; and I hate Krugman!
It's really an interesting situation; thinking about how there would no longer be an 'unambiguous' measure of time when we have faster than light travel.
It made me realise how political science can be and how facts on large issues can be covered up for political reasons.
You mean all the facts that Ruston left out made you realize that? I hope that's what you mean, as that paper is a dog.
Here's a better paper on closely related topics, by authors who have advanced the research considerably:
Nisbett, R. E., Aronson, J., Blair, C., Dickens, W., Flynn, J., Halpern, D. F., & Turkheimer, E. (2012). Intelligence: New findings and theoretical developments. American Psychologist, 67, 130-159.
The paper is not perfect, but it let me question the reality of the world around me and how much was real and how much was constructed.
I find papers that challenge my ideas more enlightening than ones that reinforce them. Which obviously makes sense I guess.
The paper I linked I considered reputable enough, in a topic known to be difficult, to be of note. Every paper has issues, the trick is working out if the issues kill the paper or not.
Perhaps not quite what the OP had in mind, but I found the papers that affected my life the most were not in my chosen profession.
Comparison of the Atkins, Zone, Ornish, and LEARN Diets for Change in Weight and Related Risk Factors Among Overweight Premenopausal Women[1] and Low-carbohydrate nutrition and metabolism[2]. After spending most of my life obese, even after having bariatric surgery to "correct" it, I found I had to dive into the science on my own to see past the charlatans and the demagogues. These two papers lit the way for me.
I'm not going to mention a specific paper, but Papers We Love (https://github.com/papers-we-love/papers-we-love) has some good stuff on it, and the meetups have always been interesting (at least for my local chapter).
If we define favorite as most often reached for in reference to present discussions, then probably this. People are persistently surprised by the expected results of cumulative gains in all areas of life, here also:
"Those who get first-generation therapies only just in time will in fact be unlikely to live more than 20–30 years more than their parents, because they will spend many frail years with a short remaining life expectancy (i.e., a high risk of imminent death), whereas those only a little younger will never get that frail and will spend rather few years even in biological middle age. Quantitatively, what this means is that if a 10% per year decline of mortality rates at all ages is achieved and sustained indefinitely, then the first 1000-year-old is probably only 5–10 years younger than the first 150-year-old."
My pleasure. Shoot me an email at jordan at birnholtz dot com if you'd like any notes or supplementary texts (or just a friendly chat about the subject).
Partly because of the fundamental importance of the paper, elucidating the structure of DNA; partly for the wonderfully understated third to last paragraph: "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material."
This paper establishes special relativity, and is remarquable for how clear it is, revolutionizing physics while using only elementary math. The first "Kinematical" part in particular does not use anything more complex mathematically than Pythagorus theorem. It is so clear that the explanations and though experiments are reproduced in all textbooks to this day; the only change is that textbooks include diagrams.
I've attempted reading this paper about half dozen time and no, it's far from "clear". The language and description of the paper requires a LOT of context and understanding of 1900s state of world. There are quite a few "companians" that can help. Here's snippet from [1]:
Modern readers turning to Einstein’s famous 1905 paper on special relativity
may not find what they expect. Its title, “On the electrodynamics of moving bodies,”
gives no inkling that it will develop an account of space and time that will topple
Newton’s system. Even its first paragraph just calls to mind an elementary experimental
result due to Faraday concerning the interaction of a magnet and conductor.
Quantum random number generation on a mobile phone (2014)[1]
A topic which seems at first rather obscure overlaps with something relatable to yield a fascinating result. The blog post [2] was especially enticing for non-specialists like myself.
Attended his defense of this, it was a ton of fun! "Defending the thesis"-slide had Joe, sword in hand, defending the thesis against a comic-styled dragon.
89 comments
[ 4.0 ms ] story [ 146 ms ] threadhttp://research.google.com/archive/gfs.html
http://research.google.com/archive/mapreduce.html
All notes and stuff can be found here: https://sites.google.com/a/cs.usfca.edu/cs-636-2011s/schedul...
I learned a ton about MapReduce and GFS. It was a great learning experience.
There have been derivative works on giving presentations, that I also particularly like: Editorial: Effective Presentations—A Must. [2]
[0] In case you don't know of him: he is the most cited living chemist, or something to this effect
[1] http://onlinelibrary.wiley.com/doi/10.1002/adma.200400767/ab...
[2] http://onlinelibrary.wiley.com/doi/10.1002/anie.201209795/ab...
Why: RPC and its ilk make a lousy model for mobile data, since mobile devices are only occasionally connected, not permanently. Similarly, in the face network and server failures, servers can be modeled as occasionally connected as well. The "replication" mindset is far more productive when dealing with those issues. The linked paper gives a broad overview of a great number of approaches to replication, and is a great way to get the lay of the land.
http://www.bitsavers.org/pdf/xerox/parc/techReports/CSL-89-1...
It's accessible, and it's a good intro to thinking about AI. The field oughta be called even more nifty algorithms.
http://people.csail.mit.edu/brooks/papers/elephants.pdf
This paper kickstarted the concept of information theory, and was hugely influential on many fields of research. Signal-to-noise ratio, the bit, information entropy, etc. are all theories and concepts presented by Shannon.
I had the exact same thought in college when I started reading a handful of well-written papers like this myself. This really set a standard that for 60+ years now academia hasn't always maintained.
I would like to share with you a few pages from the intro to my thesis which cover Shannon's channel coding theorem. There are some nice TiKZ illustrations. http://minireference.com/static/excerpts/Shannon_channel_cod... (it's not super detailed, but the definitions of all the moving parts are given)
I'm still following quantum information theory research, but more as a spectator from the sidelines. However, a couple of weeks ago I had to come back to quantum info. theory to "defend" my academic reputation. Our colleagues from TIFR found a bug in one of our papers (http://arxiv.org/abs/1111.3645v3) so my coauthor and I had to fix it. It was kind of cool to see I hadn't "lost my quantum skills" after two years of running a business. I guess, once you go quantum you never go back? :)
in case somebody is interested in more than the 15 pages provided in the excerpt, here's ivan's complete thesis:
http://arxiv.org/pdf/1208.4188v1.pdf
Some biologists may cringe (especially the Gould haters), but I don't think I've ever been so engrossed by any other scholarly paper. It is a joy to read. Very approachable for non-biologists. The papers critique of sloppy "just so" reasoning, could easily be extended to Data Scientists/Engineers/Entrepreneurs. Highly recommend!
There are other bones to pick with Gould, but these are the ones that make him impossible to read as an interested layperson without personally verifying every sentence.
Gives a good amount of insight into how academia works for mathematics, and gives a good contrast with how CS works. Don't be scared by the abstract, it's a completely non-technical paper. The academic/research culture can be more important than the results.
http://arxiv.org/abs/math.HO/9404236
http://link.springer.com/article/10.1007%2FBF03023051#page-1
However, many people here are not going to pay the money or have access through their employer, and will miss out. Free Knuth!
Completely dissociative groupoids. http://mb.math.cas.cz/mb137-1/6.html
[0] http://www.scfbm.org/content/7/1/3
https://www.princeton.edu/~pkrugman/interstellar.pdf
(12 pages, quick read)
It's really an interesting situation; thinking about how there would no longer be an 'unambiguous' measure of time when we have faster than light travel.
I try to separate academic Krugman from New York Times Krugman. It helps.
I don't. I have much respect for the man.
But his NYT writings can sometimes lean towards what might be described as "left-wing blowhardism".
It made me realise how political science can be and how facts on large issues can be covered up for political reasons.
You mean all the facts that Ruston left out made you realize that? I hope that's what you mean, as that paper is a dog.
Here's a better paper on closely related topics, by authors who have advanced the research considerably:
Nisbett, R. E., Aronson, J., Blair, C., Dickens, W., Flynn, J., Halpern, D. F., & Turkheimer, E. (2012). Intelligence: New findings and theoretical developments. American Psychologist, 67, 130-159.
doi:10.1037/a0026699
http://people.virginia.edu/~ent3c/papers2/Articles%20for%20O...
I find papers that challenge my ideas more enlightening than ones that reinforce them. Which obviously makes sense I guess.
The paper I linked I considered reputable enough, in a topic known to be difficult, to be of note. Every paper has issues, the trick is working out if the issues kill the paper or not.
Comparison of the Atkins, Zone, Ornish, and LEARN Diets for Change in Weight and Related Risk Factors Among Overweight Premenopausal Women[1] and Low-carbohydrate nutrition and metabolism[2]. After spending most of my life obese, even after having bariatric surgery to "correct" it, I found I had to dive into the science on my own to see past the charlatans and the demagogues. These two papers lit the way for me.
1. http://jama.jamanetwork.com/article.aspx?articleid=205916
2. http://ajcn.nutrition.org/content/86/2/276.full
This paper achieves a wonderful balance between being incredibly important and almost absurdly easy to read and understand.
http://dx.doi.org/10.1371%2Fjournal.pbio.0020187
"Those who get first-generation therapies only just in time will in fact be unlikely to live more than 20–30 years more than their parents, because they will spend many frail years with a short remaining life expectancy (i.e., a high risk of imminent death), whereas those only a little younger will never get that frail and will spend rather few years even in biological middle age. Quantitatively, what this means is that if a 10% per year decline of mortality rates at all ages is achieved and sustained indefinitely, then the first 1000-year-old is probably only 5–10 years younger than the first 150-year-old."
A study of competitive swimmers and what separates the mediocre from the great, but widely applicable to many forms of excellence or greatness.
Uncertainty and the Welfare Economics of Medical Care by Kenneth Arrow (1963) [1].
This paper effectively makes the case that medical care shouldn't be treated like other goods.
If you're remotely interested in health econ/health industry, I recommend reading it.
1. https://www.aeaweb.org/aer/top20/53.5.941-973.pdf
http://www.nature.com/physics/looking-back/crick/index.html
Partly because of the fundamental importance of the paper, elucidating the structure of DNA; partly for the wonderfully understated third to last paragraph: "It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material."
http://ganapathymani.com/On%20the%20electrodymics%20of%20mov...
This paper establishes special relativity, and is remarquable for how clear it is, revolutionizing physics while using only elementary math. The first "Kinematical" part in particular does not use anything more complex mathematically than Pythagorus theorem. It is so clear that the explanations and though experiments are reproduced in all textbooks to this day; the only change is that textbooks include diagrams.
Modern readers turning to Einstein’s famous 1905 paper on special relativity may not find what they expect. Its title, “On the electrodynamics of moving bodies,” gives no inkling that it will develop an account of space and time that will topple Newton’s system. Even its first paragraph just calls to mind an elementary experimental result due to Faraday concerning the interaction of a magnet and conductor.
[1] http://www.pitt.edu/~jdnorton/papers/companion.pdf
It gives insight into the nexus between politics and student unions/student bodies in India. It reads more like a story than a scholarly article.
http://www.jstor.org/discover/10.2307/2643482?uid=3739536&ui...
A topic which seems at first rather obscure overlaps with something relatable to yield a fascinating result. The blog post [2] was especially enticing for non-specialists like myself.
1. http://arxiv.org/abs/1405.0435
2. https://medium.com/the-physics-arxiv-blog/quantum-random-num...