Inferring Morphological Rules from Small Examples using 0/1 Linear Programming

Paper i proceeding, 2019

We show how to express the problem of finding an optimal morpheme segmentation from a set of labelled words as a 0/1 linear programming problem, and how to build on this to analyse a language’s morphology. The result is an automatic method for segmentation and labelling that works well even when there is very little training data available.