arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1607.01426cs.CL
收藏

基于循环神经网络的实体、关系和文本推理链

Chains of Reasoning over Entities, Relations, and Text using Recurrent Neural Networks

College of Information and Computer Sciences(信息与计算机科学学院);University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)

Rajarshi Das, Arvind Neelakantan, David Belanger, Andrew McCallum

摘要:

我们的目标是将符号逻辑推理丰富的多步推理与神经网络的泛化能力结合起来。我们对文本和大规模知识库(KBs)中的实体和关系的复杂推理特别感兴趣。Neelakantan et al.(2015)使用RNN来构建KB中多跳路径的分布式语义;然而,由于多种原因,该方法缺乏准确性和实用性。本文提出了三个重要的建模进展:(1)我们学会了联合推理关系,实体和实体类型;(2)我们使用神经注意力建模来整合多个路径;(3)我们学会了在一个单一的RNN中共享力量,该RNN代表所有关系的逻辑组成。在一个大规模的Freebase+ Cocktail Web预测任务,我们实现了25%的错误减少,和53%的错误减少稀疏的关系,由于共享的力量。在WordNet中的推理链上,我们将平均分位数的错误减少了84%,与以前的最先进技术相比。代码和数据可在https://rajarshd.github.io/ChainsofReasoning上获得

英文摘要:

Our goal is to combine the rich multistep inference of symbolic logical reasoning with the generalization capabilities of neural networks. We are particularly interested in complex reasoning about entities and relations in text and large-scale knowledge bases (KBs). Neelakantan et al. (2015) use RNNs to compose the distributed semantics of multi-hop paths in KBs; however for multiple reasons, the approach lacks accuracy and practicality. This paper proposes three significant modeling advances: (1) we learn to jointly reason about relations, entities, and entity-types; (2) we use neural attention modeling to incorporate multiple paths; (3) we learn to share strength in a single RNN that represents logical composition across all relations. On a largescale Freebase+ClueWeb prediction task, we achieve 25% error reduction, and a 53% error reduction on sparse relations due to shared strength. On chains of reasoning in WordNet we reduce error in mean quantile by 84% versus previous state-of-the-art. The code and data are available at https://rajarshd.github.io/ChainsofReasoning

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑