Hi,
I am writing this post to share what I hope might be an interesting project related to ErgoAI, and Flora-2.
The project is called Treebank Semantics. It offers a way to process constituency tree annotations of natural language and return logic based meaning representations. The system has Flora-2 object frames as an available output format.
The project is associated with a parsed corpus of English with hand-checked analysis for approaching half-a-million words. Parsed data and further results of analysis, including Flora-2 analysis, are accessible through a web based interface. Flora-2 analysis (potentially for multiple sentences with anaphoric connections) is calculated on-the-fly from the parsed data when the "flora2" output option of the interface is chosen. A "source" view shows the exact form of the base data.
Including a treebank parser for English gives a full pipeline from raw English input to Flora-2 object frames. There is an on-line demo to illustrate such a pipeline.
The homepage for the project is:
http://www.compling.jp/ajb129/ts.htmlThe homepage for the corpus is:
http://www.compling.jp/ajb129/tspc.htmlWith thanks,
-Alastair