Hacker News new | past | comments | ask | show | jobs | submit login
Facebook Publications (facebook.com)
179 points by fizwhiz on Nov 10, 2013 | hide | past | favorite | 24 comments

Lets not forget the incredible articles in the Tandem and Digital archives. Including the famous 5-minute rule paper.

HP, Digital, Tandem archives http://www.hpl.hp.com/hpjournal/journal.html

5-minute rule paper http://www.hpl.hp.com/techreports/tandem/TR-86.1.pdf

I would love to see more companies opening up technical research documents (regardless of how formal they are).

As academia becomes more stagnated it makes sense for companies to open up and spread knowledge to drive their particular ecosystem forward.

Not to mention it's probably an excellent recruiting tool.

A lot of the big tech companies do this. Yahoo and Microsoft would be particularly active in terms of publications. Google are more secretive but do publish. IBM obviously publish a lot as well. Others that come to mind Ebay , Linkedin, HP / Compaq (Weh), various big online chinese companies, and so on.

http://labs.yahoo.com http://research.microsoft.com http://research.google.com

Microsoft research is especially amazing. Not only do they publish papers on their systems like Google and Facebook, they also do a lot of basic science on a very diverse set of topics.

The research talks are also worth checking out: http://research.microsoft.com/apps/catalog/default.aspx?t=vi...

The latest talks range from "Culture differences between US and China" to "Distributed Optimization via Alternating Direction Method of Multipliers".

Hmm... interesting. I was about to make a comment about how IBM had moved a lot of their stuff behind the IEEE paywall, as a couple of weeks ago I went looking for some IBM publication and could only find it in IEEE Explore, and something I read led me to believe that all issues of IBM Systems Journal and IBM Journal of Research and Development were now only to be found at IEEE. But I looked again just now, and it appears that most (maybe all) issues of both journals are, in fact, still freely available. Weird.

Not sure if I saw an old page, or if something has changed in the past couple of weeks, or what, but at least for now, it appears that a significant body of IBM research material is still freely available.

And to make it even more confusing, I now see that some links from http://researchweb.watson.ibm.com/journal/ redirect to the IEEE site, but the actual article is still freely downloadable. Not sure how universal that is, as the article I was trying to access a couple of weeks ago, was very much not available for free (at the time). Weird. shrug

Regardless of what IEEE or IBM are doing, in academia those who know what open access actually is support it, and those who don't have mostly been disregarding closed-access rules made by publishers.

Tools like CiteSeerX can help you find archived copies of papers around the Internet, which may mean that you will not have to contend with paywalls at all to read the papers you want. For example, Tomasulo's famous paper about out-of-order execution originally published in the IBM J. Research has an entry here http://citeseerx.ist.psu.edu/viewdoc/summary?doi= and you can note that it is available on several academics' personal sites, presumably uploaded originally for pedagogical purposes. As well, CiteSeerX keeps a cache of quite a few citations locally.

Is academia really becoming more stagnated? What do you mean by that? From where I'm standing it seems like we're moving pretty quickly.

It's not at all. Perhaps he is referring to the dwindling research dollars in the US, but the opposite is occurring in Europe and China. I'm a graduate student and most older tenured faculty have seen worse funding in the past, and hold on to the hope (for their field more than anything) that science funding in the US will remain cyclical and that this current storm shall pass.

Pixar also publishes some of their investigations - http://www.pixar.com/companyinfo/research/deb/

Hate to bring Twitter up here. But this makes one even more puzzled about where Twitter's R&D costs of $90 million goes! Sure there are open source stuff they put up on http://twitter.github.io/ and there's bootstratp. But none of it is the sort of cutting edge research that the likes of LinkedIN and FB seem to be doing. Looking up posts on their blog with "research" tag also returns dismal results.

There's a tax credit for claiming your work is R&D:


This creates an incentive for companies to find ways to report their work as R&D. (Not implying that Twitter is defrauding or that their work isn't R&D, just stating why it might be useful for a company to "invest in" R&D, especially if that is work they wanted done anyway.)

r&d most likely relates to product development and isn't academic research. By listing it as r&d they can probably write something off with the tax man.

Things such as their discovery tab, targeted tweets, trending, search and user search most likely involve (Or involved) reasonable investigation.

I've no idea what source of R&D Twitter does, but you can't necessarily judge them based on what they blog about or release papers on, plenty could be going on behind closed doors.

I was excited at first because I thought Facebook was building a social network for researchers to speed up the scientific process/peer review. Then I visited the site. Oh well.

I'm building a social writing site for researchers and scientists using Markdown syntax and LaTeX at markdawn.com — someone might be interested if they see this news.

I think research in big companies is an interesting activity, both for the companies themselves and for the research world that can receive input from "real" world problems.

I was in research before (IT security) and I often found interesting results from companies such as IBM, Intel, Microsoft and Google.

What I think is kind of strange here, is that there's not a section on security/privacy in the page... well, but perhaps I'm biased ;)

This looks like an intelligent initiative from Facebook and proves it is well on its way towards transparency and would love to see a lot more of papers/publications coming up about their tech. This would largely benefit the academia.

In my view, I love Microsoft Research.


Dealing with how families interact on Facebook looks to be particularly interesting.

I'm surprised by how relatively little machine learning research they have. Microsoft, Google, IBM and Yahoo seem much better represented at the core ML conferences like ICML and NIPS.

For a second there, I thought Facebook was going to take on the academic publishers.

Can someone download them and put nicely on one page?

Not what I hoped, but at least you don't need to log in.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact