Home

Welcome to the UCCAApp wiki!

• UCCAApp is undergoing final stages of development. A live demo for the current version can be found here.
• All of UCCA's resources, including annotated corpora and software packages, are freely available here.
• A stable, pilot version (albeit more stable) of the application can be found here.

UCCAApp is an open-source, flexible web-application for syntactic and semantic phrase-based annotation in general, and for UCCA annotation in particular. UCCAApp supports a variety of formal properties that have proven useful for syntactic and semantic representation, such as discontiguous phrases, multiple parents and empty elements, making it useful to a variety of other annotation schemes with similar formal properties. UCCAApp’s user interface is intuitive and simple, so as to support annotation by users with no background in linguistics or formal representation. Indeed, a pilot version of the application has been successfully used in the compilation of the UCCA Wikipedia treebank by annotators with no previous linguistic training.

UCCA (Universal Conceptual Cognitive Annotation) is a cross-linguistically applicable semantic representation scheme, building on the Basic Linguistic Theory typological framework and Cognitive Linguistics literature. It has demonstrated applicability to multiple languages, including English, French, German and Czech, support for rapid annotation, accessibility to non-expert annotators and stability under translation. The scheme has recently proven useful for machine translation evaluation.

In order to facilitate the adoption of UCCAApp by other research groups, we built UCCAApp using recent, standard web technology. The server is based on Django, accompanied with a PostgreSQL database. The API follows the apigee style guide, and the client is based on AngularJS.