Abstract Meaning Representation

We joined the initiative to build a formal Abstract Meaning Representation (AMR, http://amr.isi.edu/) for sentences in natural languages.

We build upon our previous expertise, experience and data such as:

• the tectogrammatical layer
• the process of annotation of Prague dependency treebanks, mainly the parallel Prague Czech-English Dependency Treebank
• our lexicographic projects, primarily PDT-Vallex

To date, our efforts include:

• adapting the annotation guidelines for Czech
• annotating a small set of sentences in Czech; the same sentences are also available in English and Chinese
• getting reading for the summer JHU workshop in Prague (July 2014)