Towards Generalized Open Information Extraction
This addresses the domain-independence challenge in OpenIE for NLP applications, though it is incremental as it builds on existing OpenIE frameworks.
The paper tackles the problem of Open Information Extraction (OpenIE) models failing to generalize across domains by proposing a new benchmark and method, resulting in DragonIE outperforming previous methods by up to 6.0% in F1 score in both in-domain and out-of-domain settings.
Open Information Extraction (OpenIE) facilitates the open-domain discovery of textual facts. However, the prevailing solutions evaluate OpenIE models on in-domain test sets aside from the training corpus, which certainly violates the initial task principle of domain-independence. In this paper, we propose to advance OpenIE towards a more realistic scenario: generalizing over unseen target domains with different data distributions from the source training domains, termed Generalized OpenIE. For this purpose, we first introduce GLOBE, a large-scale human-annotated multi-domain OpenIE benchmark, to examine the robustness of recent OpenIE models to domain shifts, and the relative performance degradation of up to 70% implies the challenges of generalized OpenIE. Then, we propose DragonIE, which explores a minimalist graph expression of textual fact: directed acyclic graph, to improve the OpenIE generalization. Extensive experiments demonstrate that DragonIE beats the previous methods in both in-domain and out-of-domain settings by as much as 6.0% in F1 score absolutely, but there is still ample room for improvement.