‘(It) could potentially be used by journalists to help communicate complex research to the public, though the authors say they aren’t going to be putting journalists out of a job any time soon.‘
A new algorithm for summarizing text could make piles of information easier to absorb, reports Inside Higher Ed. Distilling texts is complicated for machines because, typically, they can’t retain context over long passages.
This new algorithm uses a new approach.
It looks for patterns between words in a document and knits them together. Researchers call their method RUM, for Rotational Unit of Memory. The name describes how meaning is drawn around a document, like how a sweep hand works on a clock.
The research was done at MIT.
OUR TAKE
- Unlike many systems that use a kind of ‘cut-and-paste’ approach that reduces the number of existing words, RUM appears to generate new sentences while retaining the original meaning.
- These are preliminary results and only indicate new research directions.
- Future iterations could be valuable for journalism. Think of applications in radio newscasts, digital signage, and aspects of daily news.
- Many forms of automated summarization today are limited to shorter documents or structured data.
Drowning in Research Reading? AI Could Help
INSIDE HIGHER ED | May 14, 2019 | by Lindsay McKenzie