ESA: Entity Summarization with Attention

TL;DR

This paper introduces ESA, a neural network model with supervised attention mechanisms designed to improve entity summarization from knowledge graphs, outperforming traditional methods in summary quality.

Contribution

The paper presents a novel deep learning approach with supervised attention for entity summarization, addressing previous limitations of clustering and graph-based techniques.

Findings

ESA improves F-measure and MAP over state-of-the-art methods

Supervised attention effectively ranks facts for better summaries

Model demonstrates robustness across different datasets

Abstract

Entity summarization aims at creating brief but informative descriptions of entities from knowledge graphs. While previous work mostly focused on traditional techniques such as clustering algorithms and graph models, we ask how to apply deep learning methods into this task. In this paper we propose ESA, a neural network with supervised attention mechanisms for entity summarization. Specifically, we calculate attention weights for facts in each entity, and rank facts to generate reliable summaries. We explore techniques to solve difficult learning problems presented by the ESA, and demonstrate the effectiveness of our model in comparison with the state-of-the-art methods. Experimental results show that our model improves the quality of the entity summaries in both F-measure and MAP.