An important step for understanding the semantic content of text is the extraction of semantic relations between entities in natural language documents. Automatic extraction techniques have to be able to identify different versions of the same relation which usually may be expressed in a great variety of ways. Therefore these techniques benefit from taking into account many syntactic and semantic features, especially parse trees generated by automatic sentence parsers. Typed dependency parse trees are edge and node labeled parse trees whose labels and topology contains valuable semantic clues. This information can be exploited for relation extraction by the use of kernels over structured data for classification. In this paper we present new tree kernels for relation extraction over typed dependency parse trees. On a public benchmark data set we are able to demonstrate a significant improvement in terms of relation extraction quality of our new kernels over other state-of-the-art kernels.

The annotation of words and phrases by ontology concepts is extremely helpful for semantic interpretation. However many ontologies, e.g. WordNet, are too fine-grained and even human annotators often have disagreements about the precise word sense. Therefore we use coarse-grained supersenses of WordNet. We employ conditional random fields (CRFs) to predict these supersenses taking into account the interaction of neigboring words. As the annotation of training data is costly we modify the CRF algorithm to process lumped labels, i.e. a set of possible labels for each training example, one of which is the correct label. This information can be derived automatically from WordNet. By this and the use of new features we are able to increase the f-value for about 8% compared to previous results. If we use training data without annotations it turns out that the resulting F-value is only slightly lower than for a fully labeled training data. Therefore unlabeled data may be employed in an unsupervised way to increase the quality of predicted supersenses.