Optimizations

WMTrans technology is intrinsically highly optimized because it uses finite state machines. Benefits include run-time efficiency and low memory requirements. Finite state machines are proven to be the best possible implementation method for word recognition. For further reference see:

• [Koskenniemi 1983] Kimmo Koskenniemi. 1983. Two-level Morphology. A General Computational Model for Word-Form Recognition and Production. Department of General Linguistics. University of Helsinki.


• [Karttunen 1994] Lauri Karttunen. 1994. Constructing Lexical Transducers. In The Proceedings of the Fifteenth International Conference on Computational Linguistics. Coling 94, I, pages 406-411, Kyoto, Japan.