Paper: Semantic Consistency: A Local Subspace Based Method for Distant Supervised Relation Extraction

ACL ID P14-2117
Title Semantic Consistency: A Local Subspace Based Method for Distant Supervised Relation Extraction
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2014
Authors

One fundamental problem of distant supervi- sion is the noisy training corpus problem. In this paper, we propose a new distant supervi- sion method, called Semantic Consistency, which can identify reliable instances from noisy instances by inspecting whether an in- stance is located in a semantically consistent region. Specifically, we propose a semantic consistency model, which first models the lo- cal subspace around an instance as a sparse linear combination of training instances, then estimate the semantic consistency by exploit- ing the characteristics of the local subspace. Experimental results verified the effectiveness of our method.