Unsupervised Numerical Information Extraction via Exploiting Syntactic Structures

Numerical information plays an important role in various fields such as scientific, financial, social, statistics, and news. Most prior studies adopt unsupervised methods by designing complex handcrafted pattern-matching rules to extract numerical information, which can be difficult scale the open domain. Other supervised require extra time, cost, knowledge design, understand, annotate training data. To address these limitations, we propose QuantityIE, a novel approach extracting structured representations exploiting syntactic features of both constituency parsing (CP) dependency (DP). The extraction results may also serve distant supervision for zero-shot model training. Our outperforms existing from two perspectives: (1) are simple yet effective, (2) more self-contained. We further retrieval based on QuantityIE answer analytical queries. Experimental demonstrate effectiveness with high fidelity.