Shallow Processing with Unification and Typed Feature Structures --- Foundations and Applications

Witold Drozdzynski; Hans-Ulrich Krieger; Jakub Piskorski; Ulrich Schäfer; Feiyu Xu

In: KI - Künstliche Intelligenz, German Journal on Artificial Intelligence - Organ des Fachbereiches "Künstliche Intelligenz" der Gesellschaft für Informatik e.V. (KI), Vol. 1, No. 1, Pages 17-23, 2004.


We present SProUT, a platform for the development of multilingual shallow text processing systems. A grammar in SProUT consists of a set of rules, where the left-hand side is a regular expression over typed feature structures (TFSs), representing the recognition pattern, and the right-hand side a TFS, specifying how the output structure looks like. The reusable core components of SProUT are a finite-state machine toolkit, a regular compiler, a finite-state machine interpreter, a typed feature structure package, and a set of linguistic processing resources. Several applications which make use of SProUT are presented. The system is implemented in Java and C(++), and runs under both MS Windows and Linux.


Weitere Links

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence