For any particular subfield of CS (Machine learning, Computer Vision, HCI, etc.), find some top universities that offer graduate seminars, find the website which will usually list a reading list.
Completely agree with this approach. Conferences, at least for me, are not that helpful. Grad level advanced courses are usually the best resource because other people have "filtered" interesting content and organized it into some common theme e.g, I recently took Kai Li's course on "Big Data":
If you have a specific topic in mind, find out the top conference(s) in the area (for many areas in CS the top resources tend to be conferences, for other topics it would be journals), and read all of the abstracts. You'll get a feeling on what is considered 'bleeding edge'.
From that start point - just 'crawl' for (a) interesting papers cited there, and (b) other things written by authors you found interesting, they'll often have the non-paywall versions of their papers available on their site.
For the lazy person who wants to cut down on the "crawling", try this listing of Best Paper Awards across pretty much every CS discipline: http://jeffhuang.com/best_paper_awards.html
The ACL Anthology is a free online archive of the last couple of decades' worth of Association for Computational Linguistics-sponsored journals and conference proceedings, which is where many people working in computational linguistics, natural language processing, machine translation, etc. hang out and publish.
HOWTO get started:
1. Pick a recent conference;
2. Click on the titles of any articles that look interesting;
3. When you get to the end of each article, dig through its bibliography. Many/most of the paper's references will also be available in the ACL Anthology; if not, Google Scholar will probably be a good resource for chasing them down. GOTO step 2.
There are many excellent sources nowadays. Especially if you're into physics or economics but not necessarily.
The most general is the Daily Roundup at National Affairs where you get titles and abstracts of papers fitting the daily theme (lousily). RSS available.
If you don't know what to read, why do you want to read papers? What are you hoping to get out of this? The collective academic works are immense. It is hard to know how to give advice without knowing more.
If you're just starting reading academic papers, I highly suggest delving into background and seminal papers in the field first, because the terminology in contemporary papers is usually pretty precise and depends on already knowing the field. Typically my workflow looks like:
1. Pick a field.
2. Look on Wikipedia, get a feel for history and seminal papers. Also
cross-reference any summary papers that you might see in ACM/Nature/Science.
3. Flip through a textbook on the topic and see if you can start grokking the
terms. The useful textbooks will probably name more fundamental research
that will serve as good reading.
4. Start digging back in sources. You'll start seeing familiar names pop up,
these are usually the seminal authors in the field, or people who write
really good summary papers.
If you google the topic, and all the papers you find reference "x et al"
in their abstract, you probably want to find that paper.
5. Read from back to front. Skip a little in the middle if pressed for time.
6. Now you can read papers on Nature/arXiv/Google and be well-prepared for the
terminology.
I like to study synchrotrons and particle accelerators in my spare time, so I looked at Wikipedia to find out about the history of the field[1]. This lead me to Courant and Snyder's landmark paper[2] about the basis of strong focusing (how to keep all those particles in a thin ring without gigantic, building-sized magnets). Through some Googling and helpful advice I found a free textbook about accelerators[3], which lead me to the granddaddy of papers, M. Sand's summary of basically everything accelerator related [4].
I think the above approach should work for any field, but the openness of the field will vary a lot. Physics and math typically have large collections of PDFs online, but I'm not sure how good CS or bioengineering is in that regard.
There is usually very little value in individual papers. You have to read a lot of them, not entirely though. If you don't know anything about a topic, start with wikipedia, check the references, also check google and google scholar with some relevant keywords. Read some papers, read at least the introduction and the related work section. From that it's usually easy to identify what are the important references and the important keywords. After checking 5 / 10 papers you usually have a good understanding of the important problems of the topic and you can call yourself an expert :)
Most of the ACM DL papers can be found online somewhere else (like via citeseer, or Google Scholar). For moral reasons, don't support the academic pay walls if you can help it.
The best option is as many have pointed out - Google Scholar.
But the very next I have personally found is ssrn.com - which is an open source repository.
You will also find that more and more papers are being hosted on arXiv.org.
If I am reading a book that links to specific academic papers, I'll first try google scholar, then a google search for the primary author. If the first doesn't return a link to a free copy, you will usually find it on the authors .edu homepage.
I receive monthly Table of Contents [ToC] alert emails from specific journals in which I'm interested, mainly chemistry and physics. This typically involves simply creating an account at the publishers website and choosing which of their journals you'd like to receive ToC alerts for, e.g. the American Chemical Society (acs.org), and then sign up for e.g. Journal of Chemical Theory and Computation alerts.
Also, for computational chemistry, at least, there's the Computational Chemistry Highlights blog [ http://www.compchemhighlights.org/ ], which selects (what they think are) interesting papers published within the past two years or so.
Finally, I follow scientiests' blogs and social media, who will typically share excellent papers that they find. Email lists like the Computational Chemistry List [ ccl.net ] will usually discuss papers as they relate to the conversations at hand.
Unfortunately, many of the excellent chemistry journals are behind pay walls.
This is how I harvest info for chemistry. I think it's highly probably that this same method could be used for all scholarly disciplines.
Hi frigg. GuerraEarth here. You might do what I just did. I looked at your profile to see what kinds of comments you've made in the past to get an idea of your tastes before offering comment. That is what you want to do--don't just go to some site and read stuff. But thoughtfully follow your interests (your wallpaper is of Mars). I am from JPL where the Mars rovers were built. In a big sandbox. The prototypes. A play box. They build all kinds of cool stuff there. If you go to the JPL page and choose scientists that do research you like, read some of their papers. All the people at JPL are good or they aren't at JPL. And look at this really interesting new toy that let's u see the surface of the sun.
Figure out which journals cover the subject matter you're most interested in (likely ~5), and read the table of contents whenever a new issue comes out.
If you want to stay up to speed on the latest and greatest across a wide range of topics, Science and Nature are often worth a read; those are weekly.
And then once you have favorite authors, you can set up automated searches to alert you whenever they publish something.
Having a peek at a listing of the talks being presented at conferences usually gives you a sneak preview of what's going to be published "soon" since people often give talks on a subject before they publish a paper on it. Conferences are also broken into subject areas, again helping you find what's of interest to you.
While this doesn't help you decide on the topic, if you have even a vague sense for what you want to read about, most people just use Google Scholar.
Nearly all CS papers are posted by the authors on personal pages in PDF form. Once you've found something that has an interested abstract just go back to regular Google with the title, add "pdf" and it's almost certain you'll get a link to the file.
[edit]: Also, once you've got one paper that fits squarely into an area of interest you can then just start reading references. You'll also get a better handle on terminology which should make your Google Scholar searches more effective.
As a student, you can get heavily discounted rates to most journals.
* For ~300$/year, I receive by mail many journals and
transactions from ACM.
* For ~40$/year, I have access to some of IEEE Xplore.
By far, I find that ACM's papers are more interesting and I'm glad for the 300$ to get the papers in dead tree form. I'm mostly reading about Computer Systems and Artificial Intelligence. For Xplore, I read the Software magazine, it's ok.
Now I'm an undergrad and I don't understand 50% of what's written in those. There's the 50% I can understand which gives me new ideas, new tools or better understanding. For the 50% I don't grok, I know planting the seeds in my brain will eventually lead to analogies and eventual understanding. Surely it can't be bad to read more than less.
This might sound like a cop out, but mostly as sources of other articles I come across. I'm routinely shocked by how often I'll read a blog post summarizing a paper, then read the actual paper and draw a much different conclusion than the post that led me there.
Science Daily (http://www.sciencedaily.com/) is a pretty good place to scan headlines, and they link to the original sources if you wan to get to the nitty gritty.
I find the Science (http://www.sciencemag.org/rss/podcast.xml) and Nature (http://feeds.nature.com/nature/podcast/current) podcasts very useful. Apart from discussing the research they publish, with their authors, they also summarize science papers appeared elsewhere. Every week there's at least one paper I want to read in full -- I'm not saying that I do.
I'm a grad student in astronomy, and I try to read one paper per day, so I've thought a bit about where to best find good papers to read. Firstly, if you are interested in finding papers in astronomy, the best resource is the ADS search:
You can search for papers any which way: by author name, by journal, by year, or some combination. Links to the papers themselves are all on ADS. Older papers are available for free on the journals' websites, and more recent papers are available for free on arxiv.org.
When looking for good papers to start with, you can just search for all papers written in a particular range of years and sort the results by the number of times the paper has been cited. (Usually the citation count is a good estimator of the importance of the paper.) If you search for all papers written within a range of years, though, make sure to select the sort by citation count option before doing your search. (Otherwise ADS will only sort the top 200 results that it returns, which will just be the 200 authors whose names come first alphabetically.)
If you're more interested in physics than astronomy, there's also an option to include physics papers in your search.
I would recommend browsing through the most highly cited papers of the past decade or two. When you find one that interests you, ADS will also give you links to all the papers that it cites, and all the papers that have cited it. You can then sort those results by citation count to find another important, related paper. After doing this for a while you start to read through a network of papers in a particular field and get a grasp of how the field has developed and what the current state of the field is.
Another good resource to find papers is review articles. The most important journal for review articles in astronomy is Annual Reviews of Astronomy & Astrophysics (ARA&A). Browse through their recent volumes for an article that interests you, then skim the article. It will give you an idea of the state of the field and will summarize all the recent, important papers in that field. (If it's a good review article, anyway....) Some other review journals that I've used include Reviews of Modern Physics, Living Reviews in Relativity, and Space Science Reviews. (As a note, there are review journals in every field, so this technique works in disciplines other than astronomy. In fact, ARA&A is published by a group which publishes review journals in many disciplines, so you can find many other review journals on their site. I can't speak to the quality of those journals, though.)
An interesting journal that I like to look through every now and again, too, is the American Journal of Physics. It's not meant to publish new physics, per se, but is more oriented towards developing a better understanding of "solved physics." So, for example, in the current issue, there is a paper on explaining how magnetic traps work and another paper which provides a new proof of Bell's inequality. Nothing truly new there, but it can help you gain a deeper understanding of physics, and oftentimes you don't need much background in physics or math to understand the papers.
It really depend on what your topic of interest is. But This is what I do and hopefully you can apply it to any field. Identify the best academic conferences on the topic you want (eg. CCS, NDSS, ACSAC, ESORICS, USENIX Security, S&P for security conferences). Then look at the abstract on the conference website and find what you think is interesting and then find the article on the author webpages or in the conference proceedings (online for usenix other you will find in ACM Digital library or IEEE Explorer).
A bunch of prominent people in computational biology are on Twitter. I follow them and between their tweets and things they retweet a bunch of interesting papers show up.
Conferences and surveys. Expect the firsts to contain a lot of noise and the second ones to be quite outdated, so search for the papers in google scholar and surf by back-references.
Imho, you should spend half a day or so going through 20-50 papers, reading abstracts+intros+conclusions to figure out which are the most interesting and select the 2 candidates. Especially at the beginning, focus on good/known authors.
BTW, surveys are also quite interesting papers by themselves.
Do you want to go broad or deep? One strategy is to find your first paper (or book with references) and then go through each of the papers it cites, and each of the papers those cite, and so on. A lot of the papers probably won't be interesting but it can be an efficient way to find quintessential papers in a field that get cited a lot.
A tip: Check if your local public library offers online access to academic journals. Mine offers access to MasterFile Premier, Academic Search Complete, GreenFile, ERIC, Medline, and a bunch of others. I used this (free) access to do the majority of the research for "Experimenting With Babies."
I recently got a bunch from sillysaurus2 (https://news.ycombinator.com/item?id=6345990). I personally feel like he has great taste, so he's a pretty good filter for interesting papers.
For example, from UW:
Machine Learning: http://courses.cs.washington.edu/courses/cse590m2/09au/ HCI: http://courses.cs.washington.edu/courses/cse590h/13au/
I've found graduate reading seminars usually try to present a mix of seminal, eclectic and recent that gives you a diverse overview of the field.