Using AI to determine how "interesting" citations are
Location
Ngee Ann Kongsi Auditorium (NAKA)
Start Date
4-6-2026 10:30 AM
End Date
4-6-2026 11:00 AM
Description
It's now easier than ever to collect and count academic or policy citations to a piece of scholarship. That makes quantitative analysis easier, but if we want to use more qualitiative assessment where we can - and generally we should - then we're faced with a lot of manual analysis that requires both contextual knowledge and a broader understanding of 'impact'. In this presentation we'll present some experiments we've done with qualitative assessment of policy impact using LLMs and other machine learning approaches, and discuss what the opportunities and challenges are for the future.
Using AI to determine how "interesting" citations are
Ngee Ann Kongsi Auditorium (NAKA)
It's now easier than ever to collect and count academic or policy citations to a piece of scholarship. That makes quantitative analysis easier, but if we want to use more qualitiative assessment where we can - and generally we should - then we're faced with a lot of manual analysis that requires both contextual knowledge and a broader understanding of 'impact'. In this presentation we'll present some experiments we've done with qualitative assessment of policy impact using LLMs and other machine learning approaches, and discuss what the opportunities and challenges are for the future.