Beyond Templates and BERT: Headword-centric parsing for semantic question answering in non-english financial domains
Jamal Al Qundus, Bassam Al-Shargabi · 2026
Read Beyond Templates and BERT: Headword-centric parsing for semantic question answering in non-english financial domains by Jamal Al Qundus, Bassam Al-Shargabi (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-0347261.txt.
Description
Jamal Al Qundus and Bassam Al-Shargabi built a question-answering system for German financial language, where long compound words and flexible phrasing defeat many artificial intelligence tools. Their 2026 system, called INAGQA, works by finding the headword of each question, the core noun everything else hangs on, then matching it against a knowledge graph of company facts.
Tested on more than two thousand expert-checked questions, the system beat both a leading language-model approach and template-based rivals, and handled time- and quantity-flavoured questions almost flawlessly. Analysts in a case study got answers in about two seconds. The authors propose the design as a template for other non-English specialist domains.
Catalog record
| CREATOR | Jamal Al Qundus, Bassam Al-Shargabi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56274 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |