A Modular Unsupervised Framework for Attribute Recognition from Unstructured Text
This addresses the need for adaptable attribute recognition in domains like incident reports, but it is incremental as it combines existing lexical and semantic similarity techniques.
The authors tackled the problem of extracting structured attributes from unstructured text without task-specific fine-tuning, achieving effective attribute extraction on a missing person use case using the InciText dataset.
We propose POSID, a modular, lightweight and on-demand framework for extracting structured attribute-based properties from unstructured text without task-specific fine-tuning. While the method is designed to be adaptable across domains, in this work, we evaluate it on human attribute recognition in incident reports. POSID combines lexical and semantic similarity techniques to identify relevant sentences and extract attributes. We demonstrate its effectiveness on a missing person use case using the InciText dataset, achieving effective attribute extraction without supervised training.