# MORSE: Semantic-ally Drive-n MORpheme SEgment-er

## TL;DR

This paper introduces MORSE, a morpheme segmentation framework that leverages morpho-syntactic regularities from word representations alongside orthographic features, achieving state-of-the-art results by utilizing vocabulary-wide semantic information.

## Contribution

It is the first to incorporate syntactico-semantic information from word embeddings into morpheme segmentation and provides a new dataset based on compositionality for evaluation.

## Key findings

- Achieves state-of-the-art segmentation accuracy
- Utilizes vocabulary-wide semantic regularities
- Introduces a new compositionality-based dataset

## Abstract

We present in this paper a novel framework for morpheme segmentation which uses the morpho-syntactic regularities preserved by word representations, in addition to orthographic features, to segment words into morphemes. This framework is the first to consider vocabulary-wide syntactico-semantic information for this task. We also analyze the deficiencies of available benchmarking datasets and introduce our own dataset that was created on the basis of compositionality. We validate our algorithm across datasets and present state-of-the-art results.