Relation Adversarial Network for Low Resource Knowledge Graph Completion
arXiv:1911.03091
Abstract
Knowledge Graph Completion (KGC) has been proposed to improve Knowledge Graphs by filling in missing connections via link prediction or relation extraction. One of the main difficulties for KGC is a low resource problem. Previous approaches assume sufficient training triples to learn versatile vectors for entities and relations, or a satisfactory number of labeled sentences to train a competent relation extraction model. However, low resource relations are very common in KGs, and those newly added relations often do not have many known samples for training. In this work, we aim at predicting new facts under a challenging setting where only limited training instances are available. We propose a general framework called Weighted Relation Adversarial Network, which utilizes an adversarial procedure to help adapt knowledge/features learned from high resource relations to different but related low resource relations. Specifically, the framework takes advantage of a relation discriminator to distinguish between samples from different relations, and help learn relation-invariant features more transferable from source relations to target relations. Experimental results show that the proposed approach outperforms previous methods regarding low resource settings for both link prediction and relation extraction.
arXiv commentWWW2020
References in corpus (15)
- RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
- Domain Adaptation: Learning Bounds and Algorithms
- KG-BERT: BERT for Knowledge Graph Completion
- Wasserstein Distance Guided Representation Learning for Domain Adaptation
- Relation Classification via Recurrent Neural Network
- Analogical Inference for Multi-Relational Embeddings
- Combining Neural Networks and Log-linear Models to Improve Relation Extraction
- Partial Transfer Learning with Selective Adversarial Networks
- Improved Neural Relation Detection for Knowledge Base Question Answering
- KBGAN: Adversarial Learning for Knowledge Graph Embeddings
- DSGAN: Generative Adversarial Training for Distant Supervision Relation Extraction
- One-Shot Relational Learning for Knowledge Graphs
- Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs
- Transfer Learning for Relation Extraction via Relation-Gated Adversarial Learning
- Adversarial Domain Adaptation for Duplicate Question Detection