2018年10月30日 · In this paper, we propose a new model, which uses a KG-specific multi-layer recurrent neural network (RNN) to model triples in a KG as sequences. It outperformed several state-of-the-art KG completion models on the conventional entity prediction task for many evaluation metrics, based on two benchmark datasets and a more difficult dataset.

DSKG: Deep Sequential models for Knowledge Graphs. unpack the data.tar.gz, which includes all of three datasets. You can also directly click runDSKG.ipynb in this page to preview the results …

2018年8月24日 · 论文：DSKG：一种用于知识图谱补全的深度序列模型（DSKG: A Deep Sequential Model for Knowledge Graph Completion） 摘要：知识图谱（KG）补全的目标是填补知识图谱中缺失的事实，其中每个事实都可表示成一个形式为 (主体, 关系, 客体) 的三元组。 当前的知识图谱补全模型都只能通过三元组中的两个元素（比如主体和关系）来预测剩余第三个元素。 我们在这篇论文中提出了一种新模型，其中使用了专门针对知识图谱的多层循环神经网 …

2018年8月24日 · dskg 不仅能预测实体，而且还能预测整个三元组。为了评估 dskg 在直接预测三元组上的表现，我们构建了一个具有较大窗口的波束搜索器。另外也还有一些能够提升序列预测结果的复杂方法 [8]。

2018年12月7日 · In this paper, we propose a new model, which uses a KG-specific multi-layer recurrent neutral network (RNN) to model triples in a KG as sequences. It outperformed several state-of-the-art KG completion models on the conventional entity prediction task for many evaluation metrics, based on two benchmark datasets and a more difficult dataset.

In this paper, we propose a new model, which uses a KG-specific multi-layer recurrent neutral network (RNN) to model triples in a KG as sequences. It outperformed several state-of-the-art KG completion models on the conventional entity prediction task for many evaluation metrics, based on two benchmark datasets and a more difficult dataset.

2019年1月1日 · In this paper, we propose a new model, which uses a KG-specific multi-layer recurrent neutral network (RNN) to model triples in a KG as sequences. It outperformed several state-of-the-art KG...

为解决上述问题，本文提出了一种适于知识图谱的深度序列模型 DSKG (a deep sequential model) ，其中使用了一种新型结构的 RNN 。本文的主要贡献包括： 提出了一种新的知识图谱补全方法，通过扩展多层 RNN 将知识图谱建模成长度为 3 的序列。

In this paper, we propose a new model, which uses a KG-specific multi-layer recurrent neural network (RNN) to model triples in a KG as sequences. It outperformed several state-of-the-art KG completion models on the conventional entity prediction task for many evaluation metrics, based on two benchmark datasets and a more difficult dataset.

2018年8月24日 · dskg 不仅能预测实体，而且还能预测整个三元组。为了评估 dskg 在直接预测三元组上的表现，我们构建了一个具有较大窗口的波束搜索器。另外也还有一些能够提升序列预测结果的复杂方法 [8]。