Who we are (2) – Mendeley

My name is Mark Levy, and I’m a Senior Data Scientist at Mendeley, a company dedicated to producing software and services to help researchers of all kinds work more effectively. Mendeley now has over two million users, nearly all of whom take advantage of the Mendeley Desktop program to help keep track of references to the research literature they care about, and to organise and access their collections of PDF files of research articles. It’s really easy to import references into Mendeley Desktop from folders already on your hard drive, or from web pages such as Google Scholar search results, as well as from a host of other sources. Once you’ve imported references into your Mendeley library, you can sort them into folders, add notes and highlights, create citations as you type in a host of formats, and easily share references with colleagues.

Who we are - Mendeley

Developing ideas for future Mendeley services

What many Mendeley users are just starting to discover is that, powerful as it is, the Desktop software provides only a few of the services that we are now able to offer as a result of collecting and storing all of our users’ libraries, subject of course to strict privacy rules, in a single database, creating a huge central catalogue of research literature and enabling us at Mendeley to connect together related readers, authors and articles. We’re already making recommended reading and article search available in Mendeley Desktop, and taking part in the EEXCESS project is one of the ways in which we’re researching and developing ideas for our next generation of services.

EEXCESS is a really fun project to be part of because Mendeley is both a content provider and, thanks to having its own scientists and specialist software developers, a research partner, so our challenge is to think up ways of providing new features that might help all of the project content partners, but we also get to apply them directly to our own data. Two of the ideas that we’re particularly excited about are novel kinds of recommendation, in our case of research articles, but which can easily be applied to other domains.

New kinds of recommendation

The first of these is “just-in-time” recommendation, which will be like doing a search but without ever having to type a query! Imagine you’re creating a document, maybe writing a new research paper of your own, or updating a wiki page, or filling out a funding application: wouldn’t it be great if the system could recognise what you were typing, turn that into a search query, send it off to Mendeley and other EEXCESS content providers, and show you a set of references that you ought to be citing, without you having to do anything else at all? The other idea we’re specially looking forward to working on is “narrative path” recommendation. Suppose you’ve found an article that you’re keen to read but which requires some background knowledge to understand. Wouldn’t it be cool if we could take advantage of information about how our users have read getting on for 100 million of articles to suggest a list of things you should read, and the best order to read them in?

Mendeley has taken part in a number of other EU research projects some of which are still running, and right now I’m also working on one called ERASM which is helping us to develop new technology connecting readers with authors and other experts in their field of interests. EEXCESS is unusual though as it involves a large number of partners in relation to the size of our company: the first face to face EEXCESS meeting I attended meant learning the names of nearly as many people as work in our entire office!

Mark Levy - Senior Data Scientist at Mendeley

Mark Levy
Senior Data Scientist at Mendeley
E-Mail: mark.levy@mendeley.com

Leave a Reply

Your email address will not be published. Required fields are marked *