Paper: Semi-Supervised Learning for Relation Extraction

ACL ID I08-1005
Title Semi-Supervised Learning for Relation Extraction
Venue International Joint Conference on Natural Language Processing
Session Main Conference
Year 2008

This paper proposes a semi-supervised learn- ing method for relation extraction. Given a small amount of labeled data and a large amount of unlabeled data, it first bootstraps a moderate number of weighted support vectors via SVM through a co-training procedure with random feature projection and then applies a label propagation (LP) algorithm via the boot- strapped support vectors. Evaluation on the ACE RDC 2003 corpus shows that our method outperforms the normal LP algorithm via all the available labeled data without SVM boot- strapping. Moreover, our method can largely reduce the computational burden. This sug- gests that our proposed method can integrate the advantages of both SVM bootstrapping and label propagation.