LLM-based uncertainty assessment of social media situational signals for crisis reporting
Timothy Douglas, Roben Delos Reyes, Asanobu Kitamoto · 2026
Read LLM-based uncertainty assessment of social media situational signals for crisis reporting by Timothy Douglas, Roben Delos Reyes, Asanobu Kitamoto (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-00829.txt.
Description
Timothy Douglas, Roben Delos Reyes, and Asanobu Kitamoto add uncertainty checks to language-model reports built from social media during disasters. Posts can give fast, real-time detail about impacts and needs, but claims vary widely in quality as an event unfolds. Most systems treat every extracted claim as equally plausible, which can mislead responders working under time pressure.
Their framework first classifies posts under an established awareness scheme, then rates each claim against outside proxy data while stating its own confidence. Reports then show both what is being said and how certain it is. The team applies the method to more than two hundred thousand earthquake posts, checked against USGS PAGER impact summaries.
Catalog record
| CREATOR | Timothy Douglas, Roben Delos Reyes, Asanobu Kitamoto |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 73 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.00829