Using HMMs To Attribute Structure To Artificial Languages
Eryilmaz, K., Little, H., & De Boer, B.
Using HMMs To Attribute Structure To Artificial Languages. In S. G. Roberts, C. Cuskley, L. McCrohon, L. Barceló-Coblijn, O. Feher, & T. Verhoef (Eds.
), The Evolution of Language: Proceedings of the 11th International Conference (EVOLANG11)
. Retrieved from http://evolang.org/neworleans/papers/125.html.
We investigated the use of Hidden Markov Models (HMMs) as a way of representing repertoires of continuous signals in order to infer their building blocks. We tested the idea on a dataset from an artificial language experiment. The study demonstrates using HMMs for this purpose is viable, but also that there is a lot of room for refinement such as explicit duration modeling, incorporation of autoregressive elements and relaxing the Markovian assumption, in order to accommodate specific details.