Iris.ai, a literature-exploration tool powered by artificial intelligence, is among a bevy of new AI-based search tools offering targeted navigation of the knowledge landscape. Using a 300-to-500-word description of a researcher's problem, or the URL of an existing paper, the service returns a map of thousands of matching documents, visually grouped by topic. The results provide "a quick and nevertheless precise overview of what should be relevant to a certain research question," says computer scientist Christian Berger at the University of Gothenburg.
Whereas conventional tools act largely as citation indices, AI-based ones can offer a more penetrating view of the literature, says Giovanni Colavizza, a research data scientist at the Alan Turing Institute in London, who studies full-text analysis of scholarly publications.
AI-based 'speed-readers' are useful because the scientific literature is so vast. By one estimate, new papers are published worldwide at a rate of 1 million each year—that's one every 30 seconds.
Iris.ai groups documents into topics defined by the words they use. Iris.ai trawls the Connecting Repositories collection, a searchable database of more than 134 million open-access papers, as well as journals to which the user's library provides access. The tool blends three algorithms to create 'document fingerprints' that reflect word-usage frequencies, which are then used to rank papers according to relevance, says Iris.ai chief technology officer Viktor Botev.
View Full Article