Bottom-up unranked tree-to-graph transducers for translation into semantic graphs

We develop a finite-state transducer for translating unranked trees into general graphs. This work is motivated by recent progress in semantic parsing natural language, where sentences are first mapped tree-shaped syntactic representations, and then these translated graph representations. investigate formal properties of our tree-to-graph transducers polynomial time algorithm weighted language input packed representation, from which best-score graphs can be efficiently recovered.