Rivendell: Project-Based Scientific Paper Search

While working on the Tolkien molecular robotics project, we realized we were having trouble finding papers. Searching on PubMed, Google Scholar, Arxiv and others for relevant papers was a Sisyphean task: First, there are hundreds of thousands of relevant papers. Second, there are many duplicates to filter when you cross lists from multiple search engines. Finally, keeping up to date was nearly impossible. This frustration is shared by anyone doing science research on the internet. Too many sources, too much hay to sort through, to find the needles. We needed a better process.

Rivendell is a new kind of meta-search engine, intended specifically for scientists, and developed by scientists. Rivendell learns what you are looking for, and will improve on your search from one query to the next. As it learns, it will suggest relevant papers, and will keep you appraised of new results.

The key insight in Rivendell is that researchers work on projects, which provide context for their search queries. You define a project in Rivendell , and it will keep learning from your queries and your relevancy marks what papers are of interest to this project. Define another project, and it will do the same, separately, for the other project. This gives Rivendell an edge over general search engine. Google scholar, for example, does not learn effectively about your interests from one query to the next, because your search queries over a period of time are in different contexts. One day you might be looking for papers on multi-robot systems. The next, you might be looking for papers on plan recognition. Google does not know. Rivendell does.

Try it for yourself. Click on the link above.