Classifying Relations via Long Short Term Memory Networks along Shortest Dependency Path

Abstract

Relation classification is an important research arena in the field of natural language processing (NLP). In this paper, we present SDP-LSTM, a novel neural network to classify the relation of two entities in a sentence. Our neural architecture leverages the shortest dependency path (SDP) between two entities; multichannel recurrent neural networks, with long short term memory (LSTM) units, pick up heterogeneous information along the SDP. Our proposed model has several distinct features: (1) The shortest dependency paths retain most relevant information (to relation classification), while eliminating irrelevant words in the sentence. (2) The multichannel LSTM networks allow effective information integration from heterogeneous sources over the dependency paths. (3) A customized dropout strategy regularizes the neural network to alleviate overfitting. We test our model on the SemEval 2010 relation classification task, and achieve an $F_1$-score of 83.7\%, higher than competing methods in the literature.

EMNLP '15

References in corpus (5)

• Improving neural networks by preventing co-adaptation of feature detectors
• Learning to Execute
• Classifying Relations by Ranking with Convolutional Neural Networks
• Long Short-Term Memory Over Tree Structures
• Discriminative Neural Sentence Modeling by Tree-Based Convolution

Cited by in corpus (43)

• Relation Classification via Recurrent Neural Network
• End-to-End Relation Extraction using LSTMs on Sequences and Tree Structures
• Improved Relation Classification by Deep Recurrent Neural Networks with Data Augmentation
• Cross-Sentence N-ary Relation Extraction with Graph LSTMs
• Towards Universal Paraphrastic Sentence Embeddings
• Joint Extraction of Entities and Relations Based on a Novel Tagging Scheme
• Attention Guided Graph Convolutional Networks for Relation Extraction
• Combining Neural Networks and Log-linear Models to Improve Relation Extraction
• Natural Language Inference by Tree-Based Convolution and Heuristic Matching
• Joint Representation Learning of Text and Knowledge for Knowledge Graph Completion
• A Frustratingly Easy Approach for Entity and Relation Extraction
• Learning for Biomedical Information Extraction: Methodological Review of Recent Advances
• Reasoning with Latent Structure Refinement for Document-Level Relation Extraction
• Improving Hypernymy Detection with an Integrated Path-based and Distributional Method
• Incorporating Relation Paths in Neural Relation Extraction
• Efficient long-distance relation extraction with DG-SpanBERT
• Neural Relation Extraction via Inner-Sentence Noise Reduction and Transfer Learning
• Natural Language Generation with Neural Variational Models
• A Comparative Study on Regularization Strategies for Embedding-based Neural Networks
• Speaker-Oriented Latent Structures for Dialogue-Based Relation Extraction
• Neural Snowball for Few-Shot Relation Learning
• Connecting the Dots: Document-level Neural Relation Extraction with Edge-oriented Graphs
• Leveraging Dependency Forest for Neural Medical Relation Extraction
• Contextualised Graph Attention for Improved Relation Extraction
• SeVeN: Augmenting Word Embeddings with Unsupervised Relation Vectors
• Syntax-BERT: Improving Pre-trained Transformers with Syntax Trees
• Neural Semantic Role Labeling with Dependency Path Embeddings
• Multiple Range-Restricted Bidirectional Gated Recurrent Units with Attention for Relation Classification
• Multi-task Domain Adaptation for Sequence Tagging
• Stochastic Learning of Nonstationary Kernels for Natural Language Modeling
• Natural Language Processing for EHR-Based Computational Phenotyping
• Looking Beyond Label Noise: Shifted Label Distribution Matters in Distantly Supervised Relation Extraction
• A Survey of Syntactic-Semantic Parsing Based on Constituent and Dependency Structures
• Global Textual Relation Embedding for Relational Understanding
• Sense Perception Common Sense Relationships
• Regularized Attentive Capsule Network for Overlapped Relation Extraction
• Temporal Information Extraction for Question Answering Using Syntactic Dependencies in an LSTM-based Architecture
• Automatic Open Knowledge Acquisition via Long Short-Term Memory Networks with Feedback Negative Sampling
• REflex: Flexible Framework for Relation Extraction in Multiple Domains
• Enriched Attention for Robust Relation Extraction
• Towards Time-Aware Distant Supervision for Relation Extraction
• Indirect Supervision for Relation Extraction using Question-Answer Pairs
• Exploiting Rich Syntax for Better Knowledge Base Question Answering