Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning
This addresses the language gap in instruction-following datasets for multilingual NLP research, though it is incremental as it builds on existing methods by extending them to new languages.
The authors tackled the lack of non-English datasets for instruction fine-tuning by creating a human-curated dataset spanning 65 languages and a larger collection of 513 million instances across 114 languages through templating and translation. They contributed four open-source resources, including the Aya Dataset and Collection, and involved collaborators from 119 countries as a case study in participatory research.
Datasets are foundational to many breakthroughs in modern artificial intelligence. Many recent achievements in the space of natural language processing (NLP) can be attributed to the finetuning of pre-trained models on a diverse set of tasks that enables a large language model (LLM) to respond to instructions. Instruction fine-tuning (IFT) requires specifically constructed and annotated datasets. However, existing datasets are almost all in the English language. In this work, our primary goal is to bridge the language gap by building a human-curated instruction-following dataset spanning 65 languages. We worked with fluent speakers of languages from around the world to collect natural instances of instructions and completions. Furthermore, we create the most extensive multilingual collection to date, comprising 513 million instances through templating and translating existing datasets across 114 languages. In total, we contribute four key resources: we develop and open-source the Aya Annotation Platform, the Aya Dataset, the Aya Collection, and the Aya Evaluation Suite. The Aya initiative also serves as a valuable case study in participatory research, involving collaborators from 119 countries. We see this as a valuable framework for future research collaborations that aim to bridge gaps in resources.