Accurate Information Extraction from Research Papers using Conditional Random Fields

TLDR

The accuracy of research paper search engines is critical, yet applying CRFs to real‑world data requires further exploration. The study applies Conditional Random Fields to extract common fields from research paper headers and citations. The authors empirically explore CRF variants, testing Gaussian, exponential, and hyperbolic‑L1 priors, feature classes, and Markov orders to improve regularization. On a standard benchmark, the method achieves state‑of‑the‑art performance, reducing average F1 error by 36% and word error rate by 78% versus prior SVMs, and outperforms HMMs.

Abstract

With the increasing use of research paper search engines, such as CiteSeer, for both literature search and hiring decisions, the accuracy of such systems is of paramount importance. This paper employs Conditional Random Fields (CRFs) for the task of extracting various common fields from the headers and citation of research papers. The basic theory of CRFs is becoming well-understood, but best-practices for applying them to real-world data requires additional exploration. This paper makes an empirical exploration of several factors, including variations on Gaussian, exponential and hyperbolic-L1 priors for improved regularization, and several classes of features and Markov order. On a standard benchmark data set, we achieve new state-of-the-art performance, reducing error in average F1 by 36%, and word error rate by 78% in comparison with the previous best SVM results. Accuracy compares even more favorably against HMMs.