Mendeley to turn off CiteULike synchronization.
Mendeley have decided to discontinue the synchronization between their service and CiteULike, from the end of January 2013.
Using CiteULike to locate quality users
Nice tutorial
Screencast: Multiple File Attachments
For best quality, we recommend selecting 720p/fullscreen or viewing this in HD on the YouTube site.
Live synchronization of your CiteULike library to Delicious.com
On your MyCiteULike menu there’s a new item “Delicious Sync” which you can use to set up a live synchronization of your library. Initially this is restricted to “new” YahooID accounts which can be accessed without the need for CiteULike to know your delicious password, but we can do that if there’s demand.
Let us know how you get on and any enhancements you’d like.
Using CiteULike (YouTube)
A quick video guide to using CiteULike and RSS feeds to filter through articles quickly
CiteULike and DeepDyve Partner to Streamline Access to Scholarly Journals
DeepDyve and CiteULike announced today that the companies are collaborating to deliver a superior way to easily and affordably share and read scholarly information on the Internet. CiteULike’s web-based service is widely used in academic and professional circles as a way to store, organize and share scholarly papers. Through its partnership with DeepDyve, CiteULike now offers its users a simple way to rent and read the journal articles they discover for as little as $0.99.
See the press release for more details.
Data from CiteULike’s new article recommender
CiteULike’s automated recommender system, as described in Toine’s excellent post, has been running live for 6 weeks now. We have some early data to share.
When users look at their recommendations list, they have the option of accepting (and thereby copying the recommended article into their own library) or rejecting (and clearing the article to make way for a new recommendation).
Here is breakdown of accepted recommendations:
and the rejected ones:
The totals are: 9930 rejected, 2323 accepted.
We are pretty excited by that percentage. It suggests that recommendations based on an algorithm working on item co-occurrence in CiteULike user’s libraries works pretty well in a real world setting. It is clearly helping people discover articles they were unaware of.
There are a huge number of refinements we can make to this, and we are busy doing just that. For example, we have not begun to look at the 6m+ tags as part of the system yet.
More than anything, I believe that it shows the quality of the citeulike dataset for a task like this. (We already make the whole CiteULike bookmarking dataset available for download; there are dozens of independent research projects using the data in institutions around the world right now).
Of course, the automated recommender system is only a few weeks old. CiteULike’s social functions have been enabling user’s to share and discover articles from day one.
We count the end result of the social discovery by the number of articles users have copied from each other’s libraries. That happened 99,159 times in the last year, 8827 times in the last 30 days.
That is almost certainly an underestimate, it does not count the times people find an article, view it on the publisher’s site and post from there.
It’s really gratifying to see the social discovery of science generated by the simple act of keeping your references public on a Web page.
New journals added to the supported sites list
We recently added the following to the list of supported sites.
Supported sites are ones that can be posted directly using, for example, our bookmarklet.
Don’t forget using our advanced bookmarklet gives enhanced functionality
- Simultaneous posting of multiple articles from PubMed
- Posting from Mendeley, Scopus and ISI Web of Knowledge
- Posting from Google Books
- Posting of a DOI, ISBN or PubMedID selected on the current page.
See the advanced bookmarklet page for how to use this and its limitations.
Science papers that interest you
CiteULike has grown to be one of the biggest and most popular social reference management websites by helping users streamline their process of storing and managing academic references. As a result of this popularity, CiteULike has seen a rich network of connections between users, items, and tags emerge. This rich tapestry of connections has enabled a wide range of options for actively supporting academics in their scientific activities using recommendation technology.
The most recent addition to CiteULike has been article recommendations: based on a user’s historical preferences for certain scientific articles and research areas, CiteULike now automatically locates and recommends interesting articles that are new to the user. Initial responses to the recommendations have been very favourable, with acceptance rates of over 30%.
However, article recommendation is not the only possible task that can be supported on CiteULike. The figure below contains an overview of nine possible recommendation tasks that can be or already are being supported by CiteULike.
Figure 1: Nine possible recommendation tasks on CiteULike. Figure is adapted from Bogers (2009) and Clements (2007).
Each of these nine tasks takes a specific object (or set of objects) as its starting point and tries to locate related objects of the same or of a different type. The colored squares indicate the four tasks that are already supported to some degree. I will briefly describe each of the nine tasks below, list in order of perceived importance and potential impact on usage of CiteULike. Ofcourse, this is completely dependent on the users of CiteULike, so comments and suggestions are most welcome!
Article recommendation The newest feature to be implemented on CiteULike is article recommendation, i.e., finding interesting, new articles for users based on the articles already in their profile. While the current recommender system seems to be performing quite well, future extensions should include better personalization and the ability to distinguish different types of article recommendation, based on the user’s information seeking needs. While recommending articles based on all articles present in a user’s reference library is an obvious and necessary first step, this is not a one-size-fits-all solution. Depending on their context and their current work tasks, users will have different goals when requesting recommendations (McNee, 2006). Examples could include (1) filling out the reference list for a paper the user is writing, (2) maintaining awareness in a specific research field, (3) finding a starting point for research, (4) exploring a research interest, and (5) exploring the novelty of a given paper.
More like this This recommendation task is related to ‘standard’ article recommendation. Instead of using the entire user profile, the user selects a single item and requests recommendations for that item. This is similar to the ‘More like this’ option by Google, where the user can ask for Web pages similar to the one the user selected. This is likely to be the next type of functionality to be added to CiteULike. [Kevin, you might to remove this if you don’t want to give it away?]
People like me The third task CiteULike could support is helping users locate other users with similar tastes. Here we start with the profile of a given users and use that to locate other, similar users. Such similar users are often to referred to as nearest neighbors. By browsing the profiles of their neighbors or by subscribing themselves to the article feeds of those users, a ‘People like me’ feature could aid in the discovery of relevant content. This feature is currently already implemented on CiteULike based on article overlap in the ‘Neighbours’ pane. In the future, this feature could be improved and expanded to include groups: a user could ask the system to recommend groups to join.
Tag suggestion Suggesting related tags when users post or copy a new article to their profile has been a popular research problem in academics with a large body of related work. Important questions here are (1) where should the recommended tags come from, (2) how do we deal with orthographic differences between similar tags, and (3) how does tag suggestion affect the user’s tagging behavior? Despite the popularity of tag suggestion as a research task, these questions have not yet been answered conclusively. Tag suggestion is already partly supported by CiteULike.
Depth browsing Depth browsing is related to the task of tag suggestion. Users select one or more tags and receive recommendations for related tags to help them in their browsing process. Supporting deeper browsing behavior is contingent on successfully identifying related tags, such as synonyms, meronyms and hyponyms, while at the same time separating the different senses of polysemous tags such as java.
Personalized search CiteULike currently already supports searching for users, articles, tags, and most of the available metadata fields. Personalizing the search process, however, could improve the quality of the results for individual users. In ‘Personalized search’, the search engine incorporates the interests of the user when deciding which content is relevant for the user. A typical way of doing this is by re-ranking the search results for a specific user using his tags.
Domain experts A seventh recommendation task that CiteULike could support is helping users locate other domain experts, i.e., users who actively use one or more specific tags to annotate their references. This could help a user locate domain experts on a specific topic by selecting a set of related tags.
People profiling We can also turn the task of finding domain experts around and generate interest or expertise profiles of users. Given a certain user, we can use the tags assigned by that user to produce an interest or expertise profile. This task is already supported by CiteULike in the form of the tag (and author clouds) present on a user’s profile page.
Item experts Locating item experts is similar to locating domain experts: who are the experts on one or more specific items? The larger the set of selected items, the easier it becomes to pinpoint the biggest experts.
References
T. Bogers, “Recommender Systems for Social Bookmarking”. Ph.D. thesis, Tilburg University, December 2009.
M. Clements, “Personalization of Social Media,” in Proceedings of the BCS IRSG Symposium: Future Directions in Information Access 2007, August 2007.
S. McNee, N. Kapoor, and J. A. Konstan, “Don’t Look Stupid: Avoiding Pitfalls when Recommending Research Papers,” in CSCW ’06: Proceedings of the 2006 ACM Conference on Computer Supported Cooperative Work, (New York, NY, USA), pp. 171–180, ACM, 2006.
Each of these nine tasks takes a specific object (or set of objects) as its starting point and tries to locate related objects of the same or of a different type. The colored squares indicate the four tasks that are already supported to some degree. I will briefly describe each of the nine tasks below, list in order of perceived importance and potential impact on usage of CiteULike. Of course, this is completely dependent on the users of CiteULike, so comments and suggestions are most welcome!
Article recommendation The newest feature to be implemented on CiteULike is article recommendation, i.e., finding interesting, new articles for users based on the articles already in their profile. While the current recommender system seems to be performing quite well, future extensions should include better personalization and the ability to distinguish different types of article recommendation, based on the user’s information seeking needs. While recommending articles based on all articles present in a user’s reference library is an obvious and necessary first step, this is not a one-size-fits-all solution. Depending on their context and their current work tasks, users will have different goals when requesting recommendations (McNee, 2006). Examples could include (1) filling out the reference list for a paper the user is writing, (2) maintaining awareness in a specific research field, (3) finding a starting point for research, (4) exploring a research interest, and (5) exploring the novelty of a given paper.
More like this This recommendation task is related to ‘standard’ article recommendation. Instead of using the entire user profile, the user selects a single item and requests recommendations for that item. This is similar to the ‘More like this’ option by Google, where the user can ask for Web pages similar to the one the user selected. This is likely to be the next type of functionality to be added to CiteULike.
People like me The third task CiteULike could support is helping users locate other users with similar tastes. Here we start with the profile of a given users and use that to locate other, similar users. Such similar users are often to referred to as nearest neighbors. By browsing the profiles of their neighbors or by subscribing themselves to the article feeds of those users, a ‘People like me’ feature could aid in the discovery of relevant content. This feature is currently already implemented on CiteULike based on article overlap in the ‘Neighbours’ pane. In the future, this feature could be improved and expanded to include groups: a user could ask the system to recommend groups to join.
Tag suggestion Suggesting related tags when users post or copy a new article to their profile has been a popular research problem in academics with a large body of related work. Important questions here are (1) where should the recommended tags come from, (2) how do we deal with orthographic differences between similar tags, and (3) how does tag suggestion affect the user’s tagging behavior? Despite the popularity of tag suggestion as a research task, these questions have not yet been answered conclusively. Tag suggestion is already partly supported by CiteULike.
Depth browsing Depth browsing is related to the task of tag suggestion. Users select one or more tags and receive recommendations for related tags to help them in their browsing process. Supporting deeper browsing behavior is contingent on successfully identifying related tags, such as synonyms, meronyms and hyponyms, while at the same time separating the different senses of polysemous tags such as java.
Personalized search CiteULike currently already supports searching for users, articles, tags, and most of the available metadata fields. Personalizing the search process, however, could improve the quality of the results for individual users. In ‘Personalized search’, the search engine incorporates the interests of the user when deciding which content is relevant for the user. A typical way of doing this is by re-ranking the search results for a specific user using his tags.
Domain experts A seventh recommendation task that CiteULike could support is helping users locate other domain experts, i.e., users who actively use one or more specific tags to annotate their references. This could help a user locate domain experts on a specific topic by selecting a set of related tags.
People profiling We can also turn the task of finding domain experts around and generate interest or expertise profiles of users. Given a certain user, we can use the tags assigned by that user to produce an interest or expertise profile. This task is already supported by CiteULike in the form of the tag (and author clouds) present on a user’s profile page.
Item experts Locating item experts is similar to locating domain experts: who are the experts on one or more specific items? The larger the set of selected items, the easier it becomes to pinpoint the biggest experts.
Toine Bogers is in the final stages of obtaining his Ph.D. degree at the Tilburg centre for Creative Computing at Tilburg University in the Netherlands. He has recently handed in his Ph.D. thesis “Recommender Systems for Social Bookmarking”, which focuses on the problem of recommending relevant articles and bookmarks on social bookmarking websites. Toine is currently working at the Royal School of Library and Information Science in Copenhagen, Denmark, where is he investigating data fusion in IR. In addition, Toine has been assisting CiteULike in implementing recommender systems technology.
References
T. Bogers, “Recommender Systems for Social Bookmarking”. Ph.D. thesis, Tilburg University, December 2009.
M. Clements, “Personalization of Social Media,” in Proceedings of the BCS IRSG Symposium: Future Directions in Information Access 2007, August 2007.
S. McNee, N. Kapoor, and J. A. Konstan, “Don’t Look Stupid: Avoiding Pitfalls when Recommending Research Papers,” in CSCW ’06: Proceedings of the 2006 ACM Conference on Computer Supported Cooperative Work, (New York, NY, USA), pp. 171–180, ACM, 2006.