From the Medium post announcing that scite has struck a deal with Wiley to embed its contextual citation snippets (dubbed, inevitably, Smart Citations) into over a hundred Wiley journals:
Smart Citations display the context of each citation to a publication, allowing readers to see not only how many times a publication has been cited but also how it was cited. Additionally, scite has deployed a state-of-the-art deep learning model to classify the citation statement as providing supporting or disputing evidence for the cited claim.
I've been following scite since its 2019 launch: It's a small, for-profit Brooklyn-based startup that launched with a $350k investment from NY-based ff Venture Capital. The firm, curiously, has won most of its subsequent funding from federal grants, including a $1.5 million award in May 2020 from the National Institute on Drug Abuse (NIDA).
The company, in effect, is trying to monetize Henry Small's 30-year-old notion of citation context analysis (elaborated in a 1982 out-of-print book chapter).1 Scite extracts the text around a citation and then—using a machine-learning model—makes a guess on valence: Is the cite "supporting," "disputing," or a mere "mention"?
Interesting. Too bad it's a for-profit, and landing deals that, possibly, will rationalize (in both senses) strategic submission behavior and prop up the journal prestige economy.
-
Small, H. G. (1982). Citation context analysis. In B. Dervin & M. J. Voigt (Eds.), Progress in communication sciences
(pp. 287–310). Norwood, NJ: Ablex. ↩