AfriWOZ: Corpus for Exploiting Cross-Lingual Transferability for Generation of Dialogues in Low-Resource, African Languages

TL;DR

This paper introduces high-quality dialogue datasets for six African languages, investigates transfer learning with state-of-the-art models, and demonstrates promising cross-lingual transferability and human-like dialogue generation in low-resource African languages.

Contribution

It provides the first dialogue datasets for six African languages and analyzes the effectiveness of transfer learning with deep models for low-resource language dialogue generation.

Findings

Deep monolingual models learn abstractions that transfer across languages.

Five out of six languages show human-like conversational quality.

Nigerian Pidgin English exhibits the highest transferability with 78.1% human-likeness.

Abstract

Dialogue generation is an important NLP task fraught with many challenges. The challenges become more daunting for low-resource African languages. To enable the creation of dialogue agents for African languages, we contribute the first high-quality dialogue datasets for 6 African languages: Swahili, Wolof, Hausa, Nigerian Pidgin English, Kinyarwanda & Yor\`ub\'a. These datasets consist of 1,500 turns each, which we translate from a portion of the English multi-domain MultiWOZ dataset. Subsequently, we investigate & analyze the effectiveness of modelling through transfer learning by utilziing state-of-the-art (SoTA) deep monolingual models: DialoGPT and BlenderBot. We compare the models with a simple seq2seq baseline using perplexity. Besides this, we conduct human evaluation of single-turn conversations by using majority votes and measure inter-annotator agreement (IAA). We find that the hypothesis that deep monolingual models learn some abstractions that generalize across languages holds. We observe human-like conversations, to different degrees, in 5 out of the 6 languages. The language with the most transferable properties is the Nigerian Pidgin English, with a human-likeness score of 78.1%, of which 34.4% are unanimous. We freely provide the datasets and host the model checkpoints/demos on the HuggingFace hub for public access.