Labeled Morphological Segmentation with Semi-Markov Models: A Unified Approach to Morphological Processing

This paper presents labeled morphological segmentation (LMS), a unified approach to morphological processing that models the distinctions between different types of morphemes and the morphotactics of a language. The authors develop CHIPMUNK, a discriminative semi-Markov model for LMS that outperforms previous approaches on three related tasks: morphological segmentation, stemming, and morphological tag classification.