Modeling Relations and Their Mentions without Labeled Text

Abstract

Abstract. Several recent works on relation extraction have been applying the distant supervision paradigm: instead of relying on annotated text to learn how to predict relations, they employ existing knowledge bases (KBs) as source of supervision. Crucially, these approaches are trained based on the assumption that each sentence which mentions the two related entities is an expression of the given relation. Here we argue that this leads to noisy patterns that hurt precision, in particular if the knowledge base is not directly related to the text we are working with. We present a novel approach to distant supervision that can alleviate this problem based on the following two ideas: First, we use a factor graph to explicitly model the decision whether two entities are related, and the decision whether this relation is mentioned in a given sentence; second, we apply constraint-driven semi-supervision to train this model without any knowledge about which sentences express the relations in our training KB. We apply our approach to extract relations from the New York Times corpus and use Freebase as knowledge base. When compared to a state-of-the-art approach for relation extraction under distant supervision, we achieve 31% error reduction.

Context-Aware Representations for Knowledge Base Relation Extraction

“…Most of the methods (Riedel et al, 2010;Zeng et al, 2015;Lin et al, 2016) focus on predicting a single relation type based on the combined evidence from all of the occurrences of an entity pair. Hoffmann et al (2011) and Surdeanu et al (2012) assign multiple relation types to each entity pair, such that the predictions are tied to particular occurrences of the entity pair.…”

“…For relation extraction, it is crucial to be able to extract relevant features from the sentential context (Riedel et al, 2010;Zeng et al, 2015). Modern approaches focus just on the relation between the target entities and disregard other relations that might be present in the same sentence (Zeng et al, 2015;Lin et al, 2016).…”

Relation extraction with weakly supervised learning based on process-structure-property-performance reciprocity

“…The score is the of each sentence. See Appendix for the source articles. Score/LabelSentence136.5/P… the following matrix form: [11]  …234.8/P…  is the characteristic or critical whisker length, and  …  is the matrix shear strength …334.2/P… toughness () and grain … dvpwhere, is the matrix …431.0/P… cast iron has a pearlite matrix and …528.6/Pafter solution treatment , the increase of grain size was not obvious because of the heat resistance introduced by … .2) after aging … .3) grain refining , size reduction of …626.0/N solution strengthening and precipitation strengthening respectively, …, was the yield strength …724.7/N…dislocation density in lath martensite matrix due to the high content of element … 100 steel delayed the recovery process during tempering …8...23.8/P lath martensite , which benefited the impact toughness …9−13.1/P… the effect of ingot grain refinement on the mechanical properties of al profiles which are manufactured through hot working …10−14.1/N… refining the prior austenitic grain size … LONG CONTEXT … the mechanical strength and cleavage resistance …11−16.4/N… enhanced solid solution strengthening and composition homogenization is larger than …12−18.7/N… as the solution treatment temperature increases to …, the transformation … and the formation of rim o phase …13−23.4/N… during the aging treatment, the rim o phase at the margin of grains become … …”

“…The factor collection described in Section 3.1 and these factors were mapped into sentences that refer each factor in Section 3.2. Unlike in previous works [7,11], the factors were not named entities. Any noun phrase can be a factor, and factors were not predefined.…”

“…In recent years, convolutional neural network (CNN) models have surpassed feature-based models [7–11]. CNNs are a class of neural networks with convoluted neural units.…”