Logo Lanfrica

Lost in Translation: Evaluating the Performance of Five Publicly Available Machine-Translators for Luganda-English Interpretation

Domain:

natural language processing

Record type:

paper
Creator:
Ric
Publisher:
Cen
Host:
Luganda, after English, is the most popular language in Uganda (Ssentanda and Nakayiza, 2015), boasting over 20 million speakers across the globe in 2021 (New Vision, 2021). Despite its ubiquity, Luganda has been labelled a “low resource language” (Kimera et al., 2025). This project aims to evaluate the performance of five publicly available services (Bing Translator, Google Translate, Perplexity AI, ChatGPT, Google Gemini) in Luganda to English and English to Luganda Translation.

Tasks

machine translation

Languages

Ganda

Licenses

https://creativecommons.org/licenses/by/4.0/legalcode

Similar

0x0checo/Luganda-to-English-machine-translation

# Fine-tuning Multilingual NMT for a Low-Resource, Morphologically Rich Language (Luganda → English)

Explainable English-Luganda Machine Translation Models for Building Inclusive Fintech Applications

English-Luganda Dataset to Leverage Machine Translation for Building Responsible Software Systems in Fintechs

English to Luganda SMT: Ganda Noun Class Prefix Segmentation for Enriched Machine Translation

Analysis of the Cybersecurity Threats in Botswana Using Publicly Available Data

Online criminal and terrorist activities impact society at individual, organizational and national l

Evaluating Native-Speaker Preferences on Machine Translation and Post-Edits for Five African Languages