This paper presents the contribution of the Unbabel team to the WMT 2017 Shared Task on Translation Quality Estimation. We participated on the word-level and sentence-level tracks. We describe our two submitted systems: (i) STACKEDQE, a "pure" QE system, trained only on the provided training sets, which is a stacked combination of a feature-rich sequential linear model with a neural network, and (ii) FULLSTACKEDQE, which also stacks the predictions of an automatic post-editing system, trained on additional data. When evaluated on the EnglishGerman and German-English datasets, FULLSTACKEDQE achieved word-level F MULT 1 scores of 56.6% and 52.9%, and sentence-level correlation Pearson scores of 64.1% and 62.6%, respectively. Our system ranked second in both tracks, being statistically indistinguishable from the best system in the word-level track.
The Linked Data initiative brought new opportunities for building the next generation of Web applications. However, the full potential of linked data depends on how easy it is to transform data stored in relational databases into RDF triples. Recently, the W3C RDB2RDF Working Group proposed a mapping language, called R2RML, to specify mappings between relational schemas and RDF vocabularies. However, the specification of R2RML mappings is not an easy task. This paper therefore proposes a strategy to simplify the specification of R2RML mappings. The paper first introduces correspondence assertions, which provide a convenient way to manually model mappings between relational schemas and RDF vocabularies. Then, the paper describes a method to automatically generate R2RML mappings from the correspondence assertions.
In this paper, we demonstrate the RBA (R2RML By Assertion) tool which automatically generates customized R2RML mappings based on a set of semantic mappings that model the relationship between the relational database schema and a target ontology in RDF. The semantic mappings are specified by a set of correspondence assertions, which are simple to understand.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.