The Chest X-ray Imaging Dataset for Multiple Cardio-respiratory Diseases in Ethiopia (Afro Chest X-ray for short) is a project funded by the LacunaFund whose aim is to close the gap in health disparities by fostering interdisciplinary collaborations that create, expand, or aggregate labeled training and evaluation datasets.

Cardio-respiratory diseases (cardiovascular and respiratory diseases) are recognized as serious, worldwide public health concerns that have remained among the leading causes of death globally. There are not many publicly available datasets from Africa making it difficult to determine whether tools and techniques developed in other geographies are as effective in our context. In this project, we propose to create a labeled chest X-ray dataset for multiple cardio respiratory diseases in Ethiopia. We will publish the dataset as open source. We believe this dataset will stimulate researchers and practitioners in Africa and beyond to push the limits of current methods to adapt them to the African context and build assistive technologies that could empower the scarce radiologists.

With the rapid development of evaluation datasets to assess LLMs understanding across a wide range of subjects and domains, identifying a suitable language understanding benchmark has become increasingly challenging. In this work, we explore LLM evaluation challenges for low-resource language understanding and introduce ProverbEval, LLM evaluation benchmark for low-resource languages based on proverbs to focus on low-resource language understanding in culture-specific scenarios. We benchmark various LLMs and explore factors that create variability in the benchmarking process. We observed performance variances of up to 50%, depending on the order in which answer choices were presented in multiple-choice tasks. Native language proverb descriptions significantly improve tasks such as proverb generation, contributing to improved outcomes. Additionally, monolingual evaluations consistently outperformed their cross-lingual counterparts. We argue special attention must be given to the order of choices, choice of prompt language, task variability, and generation tasks when creating LLM evaluation benchmarks.