Grassroots Vol 21 No 3 | Page 34

NEWS

Drowning in the literature ? These smart software tools can help

Search engines that highlight key papers are keeping scientists up to date .

Every time Eddie Smolyansky had a few moments to himself , he tried to stay abreast of new publications in his field . But by 2016 , the computervision researcher , who is based in Tel Aviv , Israel , was receiving hundreds of automated literature recommendations per day . “ At some point , the bathroom breaks weren ’ t enough ,” he says . The recommendations were “ way too much , and impossible to keep up with ”.

Smolyansky ’ s ‘ feed fatigue ’ will be familiar to many academics . Academic alert tools , originally designed to focus attention on relevant papers , have themselves become a hindrance , flooding the inboxes of scientists worldwide .

David Matthews

Current Address : Nature Reprinted from : https :// go . nature . com / 3GhdBME
“ I haven ’ t even been reading my automated PubMed searches lately because it really is overwhelming ,” says Craig Kaplan , a biologist at the University of Pittsburgh in Pennsylvania . “ I honestly cannot keep on top of the literature .”
But change is afoot . In 2019 , Smolyansky co-founded Connected Papers , one of a new generation of visual literature-mapping and recommendation tools . Other services that promise to tame the information overload , integrating Twitter feeds and daily news as well as research , are also available .
Origin story
Instead of serving up a daily list of new articles by e-mail , Connected Papers uses a single , user-chosen ‘ origin paper ’ to build a map of related research , based partly on overlapping citations . The service recently surpassed one million users , Smolyansky says .
The maps are colour-coded by publication date , and users can toggle
Figure 1 : Illustration by The Project Twins
between ‘ prior ’, seminal , papers and later , ‘ derivative ’, works that build on them . The idea is that scientists can search for an origin paper that interests them , and see from the resulting map which recent papers have made a splash in their field , how they relate to other research , and how many citations they have accrued .
“ You do not have to sit on the hose of papers and look at every paper that comes out for fear of missing it ,” says Smolyansky . The tool is also helpful when scientists want to dive into an entirely new field , he adds , providing an overview of the essential literature .
Another visual-mapping tool is Open Knowledge Maps , a service offered by a Vienna-based not-for-profit organization of the same name . It was founded in 2015 by Peter Kraker , a former scholarly-communication researcher at Graz University of Technology in Austria .
Open Knowledge Maps creates its maps based on keywords rather than a central article and relies on text similarity and metadata to work out how papers are related . The tool arranges 100 papers in similar subfields into bubbles whose relative positions suggest similarity ; a search for articles on ‘ climate change ’, for example , might yield a related bubble about ‘ risk cognition ’.
Maps of these bubbles can be built in about 20 seconds , and users can change them to include the 100 most
33 Grassroots Vol 21 No 3 November 2021