Extracting user profile via large language models and ontologies
Pegah Safari, Mehrnoush Shamsfard · 2026
Read Extracting user profile via large language models and ontologies by Pegah Safari, Mehrnoush Shamsfard (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/plos-pone-0329934.txt.
Description
Pegah Safari and Mehrnoush Shamsfard build a system that pulls a person's basic profile, such as age, job, and hobbies, out of chat text in Persian. Such profiles power suggestions and personal assistants, but big language models handle Persian poorly. Their 2026 study tests many models and finds them tripped up by scarce training material and the language's complex structure.
The team's hybrid method mixes slot filling, learning from examples in the prompt, and reasoning over a knowledge map. It beat famous large models by wide margins, even versions fine-tuned for the task. It was also far better at catching clashes in what a user claimed. The work shows smaller, well-guided tools can outdo giants in less-served languages.
Catalog record
| CREATOR | Pegah Safari, Mehrnoush Shamsfard |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 83066 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |