Abstract
Most modern Information Extraction (IE) systems are implemented as sequential taggers and only model local dependencies. Non-local and non-sequential context is, however, a valuable source of information to improve predictions. In this paper, we introduce GraphIE, a framework that operates over a graph representing a broad set of dependencies between textual units (i.e. words or sentences). The algorithm propagates information between connected nodes through graph convolutions, generating a richer representation that can be exploited to improve word-level predictions. Evaluation on three different tasks — namely textual, social media and visual information extraction — shows that GraphIE consistently outperforms the state-of-the-art sequence tagging model by a significant margin.
Highlights
Most modern Information Extraction (IE) systems are implemented as sequential taggers
We extend the graph-based approach to validate the benefits of using other types of relations in a broader range of tasks, such as coreference in named entity recognition, followed-by link in social media, and layout structure in visual information extraction
We evaluate the model on three tasks, including two traditional IE tasks, namely textual information extraction and social media information extraction, and an under-explored task — visual information extraction
Summary
Most modern Information Extraction (IE) systems are implemented as sequential taggers. Most of the prior work looking at the non-local dependencies incorporates them by constraining the output space in a structured prediction framework (Finkel et al, 2005; Reichart and Barzilay, 2012; Hu et al, 2016). Such approaches, mostly overlook the richer set of structural relations in the input space. Designing effective features is challenging, arbitrary and time consuming, especially when the underlying structure is complex These approaches have limited capacity of capturing node interactions informed by the graph structure
Sign-in/Register to view highlights & summary 
Submitted Version (
Free)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call