Zero-Shot Morphological Discovery in Low-Resource Bantu Languages via Cross-Lingual Transfer and Unsupervised Clustering

TL;DR

This paper introduces a novel method combining cross-lingual transfer learning and unsupervised clustering to discover morphological features in low-resource Bantu languages, validated on Giriama.

Contribution

It presents a new pipeline that effectively uncovers morphological patterns and improves lemmatization accuracy in low-resource languages using transfer and clustering.

Findings

Discovered two undocumented morphological patterns in Giriama.

Achieved 78.2% lemmatization accuracy on known paradigms.

Reaches 97.3% segmentation and 86.7% lemmatization on expanded corpus.

Abstract

We present a method for discovering morphological features in low-resource Bantu languages by combining cross-lingual transfer learning with unsupervised clustering. Applied to Giriama (nyf), a language with only 91 labeled paradigms, our pipeline discovers noun class assignments for 2,455 words and identifies two previously undocumented morphological patterns: an a- prefix variant for Class 2 (vowel coalescence - the merger of two adjacent vowels - of wa-, 95.1% consistency) and a contracted k'- prefix (98.5% consistency). External validation on 444 known Giriama verb paradigms confirms 78.2% lemmatization accuracy, while a v3 corpus expansion to 19,624 words (9,014 unique lemmas) achieves 97.3% segmentation and 86.7% lemmatization rates across all major word classes. Our ensemble of transfer learning from Swahili and unsupervised clustering, combined via weighted voting, exploits complementary strengths: transfer excels at cognate detection (leveraging ~60% vocabulary overlap) while clustering discovers language-specific innovations invisible to transfer. We release all code and discovered lexicons to support morphological documentation for low-resource Bantu languages.