I’ve started trying out a new service, called Mendeley. The quickest way to describe it is a “last.fm for research;” they have a desktop client that can monitor the pdf files that you are reading, and an online presence where each user has a profile. (Read about them on their blog; my profile is here). So far, it seems that they are at a very early stage. However, the basic functionality (seeing/tagging/searching papers you read) seems quite nice. On the other hand, an obvious difficulty is that of extracting accurate meta-data from research pdf files.
The similarity between research papers and songs is quite striking. Think of it this way: songs (research papers) are made by musicians (authored by researchers), have a name (title), and are collected in albums (journals/conference proceedings). Both have a time of release; both can be tagged/described/loved/hated; both are blogged and talked about. Sometimes artists make music videos, sometimes researchers make presentations or demos.
(Ok, so at the end of the day rock stars take home millions, are famous and when not on tour around the world get peeved off about the Internet facilitating theft. PhD students are happy when there is pizza at the weekly seminar – the payoff may not be quite on par yet, but the similarity between songs and papers stands).
The main problem they seem to want to tackle is that of research discovery; again, the parallel with work in music information retrieval seems obvious. Like with music, though, there are a number of advantages to having (and using) a research-profiler. Here are some thoughts:
- Learning about myself: Although a lot of the cool stuff in recommender systems is about finding things I don’t know about (based on what I know/ have read/rated), one of the most interesting parts of these social systems has been the profiling aspect: I have changed what music I listen to based on what my profile reports, and have learned a lot about trends in my own habits. Knowing what I read will tell me what conferences I favor, what authors I know the most about (or cite often), and (more importantly) the gaps that I don’t think I’ve missed.
- Finding people who share my research interests: Rather than only uncovering papers that I have missed, I want to know/connect to people who attend similar conferences, read similar papers, and write papers I am interested in. The Twitter “follow” model would be nice: I could then find and follow researchers who I am interested in, and be notified whenever they publish something new.
- Learning about (and from) my colleagues: While finding people in my field would be neat, knowing what other people who I know (and who may be working in a completely different field) are looking at might be the path to finding serendipitous papers. For example, check out this tag cloud I made of research going on in our group, based on our research paper titles. Other than helping characterise who we are and what we do, it points out that this group of people I work with cover a lot of ground when reading, and seeing what they do may help me with my work (or at least give us something to talk about at a coffee break).
So, it seems, a system like this should serve more purposes than being a collaborative filtering engine. Then again… so should all good recommender systems.