Why aren't we systematically pooling many excellent and relevant academic articles yet?

I'm well aware of the current thread on the best textbooks for any subject, but why only share "the best" "textbooks"? I understand the desire for vetted quality, but I think we're missing an awesome opportunity here. I, for one, have a long (to me) list of fascinating academic articles I'm probably not going to read, but which I think other LessWrongers might very much like to be pointed to. For example:

  • Formalizing Trust as a Computational Concept, a CS dissertation by Stephen Marsh, 1994. From the abstract: "This thesis provides a clarification of trust. We present a formalism for trust which provides us with a tool for precise discussion. The formalism is implementable: it can be embedded in an artificial agent, enabling the agent to make trust-based decisions."
  • Mechanisms of Social Cognition, Frith and Frith, 2012, Annual Review of Psychology, which includes a spectacular table on pages 4 and 5: column 1, the broad areas considered (e.g. reward learning, tracking intentions); column 2, the mechanisms involved; column 3, the brain regions involved; and column 4, the social processes to which the mechanism and/or brain region apply.
  • Peering into the Bias Blind Spot, in which we learn that part of the reason people think themselves to be less biased than the average is because they rely on introspections while looking at behavior in others.
  • Information processing, computation, and cognition (Piccinini and Scarantino, 2009), which argues that there are a number of myths concerning these topics that stand in the way of theoretical progress, myths like 'computation is the same as information processing' or 'the Church-Turing thesis entails that cognition is computation'.
  • Review articles on operant conditioning; implementation intentions, a major new scientific approach to forming habits'; the planning fallacy (and it's "cognitive, motivational, and social origins"); and procrastination (the last one has already been mentioned here in lukeprog's main procrastination article).

We could have a database of some kind. And it could include articles that have already been well-spread here, like the seminal Kahneman & Tversky papers. Or the Singularity Institute's work and related work, such as Asimov's 3 laws of robotics and machine meta-ethics (S.L. Anderson), or the Transhumanist FAQ (Nick Bostrom).

I propose that we do so, and that we seek a balance between (a) quality and quantity and (b) organization (into topics) and simply adding, but that we also not worry too much about that right now and just start something. I envision adding article entries to some kind of simple database with (i) the bibliographic information, (ii) a link to a non-gated pdf if possible, (iii) a brief description detailing why it's bound to be very interesting to some LessWrongers, and (iv) a vote-up/down mechanism to allow fellow LessWrongers to agree that 'yes, this article title+description does seem exceptionally interesting'.

Main issue: determining the best way to share articles with each other, after determining that it's something we'd like to do.

P.S. I am unfamiliar with tagging on LW articles (having not written one before), should I add a tag/tags?

New Comment
14 comments, sorted by Click to highlight new comments since:

That's the space Mendeley and CiteULike are in, for articles, and LibraryThing or GoodReads for books. (Back when I first joined I bought a LibraryThing account for SIAI that I wanted to donate, seeded with 300+ books from its shelves, but I wasn't taken up on that.)

They're all somewhat deficient, in my view, when it comes to group operation - all are heavily biased to collecting and managing an individual's references. But they're better than nothing; and better than the Wiki, which I don't think is very suitable. Bibliographic entries are a pain to work with manually.

Having a database of references, per se, isn't so valuable. What would make such a thing valuable to me is seeing annotations from the folks on LW that I can tell are more knowledgeable than me.

The Singularity Institute already maintains a Mendeley group, but it is just for literature cited by SI publications. Another Mendeley group could be created for "Useful References for Less Wrongers" or something, but it would require a Hero to maintain it, and I'm not sure anybody wants to put in the time. It sounds like a 100+ hours project to me.

That's perhaps more properly a CFAR type mission, actually.

(Justified) fear of contributing noise rather than signal.

Difficulty of organizing these articles so that it's SO easy to browse that this becomes a place to check when looking for a book.

I, for one, have a long (to me) list of fascinating academic articles I'm probably not going to read

! This sentence deserves some consideration.

Part of my work now is reading through all the papers a previous graduate student thought were important enough to save. Some of them have been very valuable- but about half of them have been totally inappropriate for my research, and it's bothersome to have to slog through enough of them to be sure that they are actually irrelevant. This is making me more pessimistic about a general repository of papers than I probably should be.

I think we're more at the point where we could use popularizers- like lukeprog's article on procrastination- than reference lists. But popularization is far more work, and so a reference database may be a better first step.

Thanks for the ultimately encouraging comment. Agreed that there is such a great quantity of possible papers to read that some care must be taken in what one recommends. To some extent, I think we'd have to wait-and-see how conscientious/well-targeted fellow LessWrongers are in their recommendations.

[-][anonymous]40

This would be somewhat useful to me. Every day or so I read the arXiv preprint server's RSS feed, and the inhabitants of #lesswrong are no doubt familiar with me flagging articles of possible interest.

Maybe once a week you could post (to the open thread) a list of all "articles of possible interest" you discovered that week?

[-][anonymous]10

Next week I'll keep a list in my local wiki and post it on Friday.

Unless I forget, or get lazy.

How to help:

Do you know of a good website for implementing this idea? (if the idea is sufficiently clear)

Do you know if something already exists in the LessWrong community and I'm just ignorant of it?

Do you have a few particularly interesting articles you'd like to share?

Do you know of a good website for implementing this idea? (if the idea is sufficiently clear)

It seems to me that the wiki would work fine.

[-][anonymous]00

I like your idea. The easiest way to implement this would be articles on the wiki, this would offer everything except a down and up vote option.

P.S. I am unfamiliar with tagging on LW articles (having not written one before), should I add a tag/tags?

We don't use them as much as we should. This is part of our problem of poor indexing of useful articles not part of any sequence. I suggest you add the tag "scholarship".

Here's another consideration: figuring out how to better target articles to those who could use them for research. For example, a particular FAI researcher may find the first useful. This would require research profiles of some kind, of course, which is getting too complicated... unless there is already a highly-used website like Mendeley that many LWers (at least, those who do a lot of research) use?

How to help: Do you know of a good website for implementing this idea? (if the idea is sufficiently clear) Do you know if something already exists in the LessWrong community and I'm just ignorant of it? * Do you have a few particularly interesting articles you'd like to share?

[This comment is no longer endorsed by its author]Reply