Researchers often need to ask questions about how topics or stances change over time in large text collections. The new arXiv preprint introduces TL;QR, a system aimed at such temporal discursive patterns.
The abstract gives an example: a political group's stance on a topic might drift from support to criticism over several years of debates. TL;QR appears designed to make these kinds of queries possible.
The paper is currently available on arXiv under identifier 2610.15712. As a preprint, it has not yet undergone peer review.