An Unsupervised Method for Uncovering Morphological Chains

TL;DR

This paper introduces an unsupervised approach combining orthographic and semantic information to analyze morphological chains, outperforming existing systems across multiple languages.

Contribution

It presents a novel log-linear model for morphological analysis that integrates semantic and orthographic features to identify morphological chains without supervision.

Findings

Outperforms five state-of-the-art systems on Arabic, English, and Turkish.

Effectively models parent-child relations in morphological chains.

Utilizes contrastive estimation for feasible training.

Abstract

Most state-of-the-art systems today produce morphological analysis based only on orthographic patterns. In contrast, we propose a model for unsupervised morphological analysis that integrates orthographic and semantic views of words. We model word formation in terms of morphological chains, from base words to the observed words, breaking the chains into parent-child relations. We use log-linear models with morpheme and word-level features to predict possible parents, including their modifications, for each word. The limited set of candidate parents for each word render contrastive estimation feasible. Our model consistently matches or outperforms five state-of-the-art systems on Arabic, English and Turkish.