Ambiguous Entity Oriented Targeted Document Detection
Wei Shen, Haixu Wen · 2024
The documents mentioning a target entity are essential prerequisites of various applications, such as market intelligence analysis, knowledge base enrichment, fact checking and retrieval augmented generation. A simple solution to acquire these documents is to exploit search engines via querying the name of the target entity. However, the name of the target entity appearing in a returned document does not necessarily mean it really refers to the target entity due to the name ambiguity, as it may refer to another entity sharing the same name as the target entity. Thus, in this paper, we explore a new task of targeted document detection, which aims to detect those targeted documents (i.e., documents really mentioning the target entity) from the given candidate documents each of which contains an ambiguous name of the target entity. We propose GADE, a novel Graph-based framework to solve the task of tArgeted Document dEtection by leveraging both the local relevance information and the global cross-document interactions jointly. We develop a local relevance model to capture the local relevance information between the target entity and the candidate document based on a pre-trained language model. Additionally, we derive a global interaction model to unify the two categories of global cross-document interactions (i.e., similarity interaction and dissimilarity interaction) among candidate documents. Specifically, a document interaction graph is constructed based on all the candidate documents of a target entity, and a graph neural network is applied over the graph to model the global cross-document interactions via message passing mechanism. We construct four labeled datasets for this task based on Wikipedia and Web documents respectively, and a thorough experimental study shows that our framework GADE significantly outperforms all the baseline methods in terms of F1-score.