Themed Challenges to Solve Data Scarcity in Africa: A Proposition for Increasing Local Data Collection and Integration
Mubaraq Yakubu, Udunna Anazodo, Maruf Adewole, Theodore Barfoot, Tiarna Lee, Tom Vercauteren, Jonathan Shapey, Andrew King, Alexander Hammers · 2025
Read Themed Challenges to Solve Data Scarcity in Africa: A Proposition for Increasing Local Data Collection and Integration by Mubaraq Yakubu, Udunna Anazodo, Maruf Adewole, Theodore Barfoot, Tiarna Lee, Tom Vercauteren, Jonathan Shapey, Andrew King, Alexander Hammers (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-00925.txt.
Introduction
Mubaraq Yakubu and co-authors propose themed challenges to grow local medical data sets across Africa, in work from twenty twenty-five. Scarce local data and compute block AI tools in African clinics. Imports trained elsewhere carry bias that misfits local patients. The gap caps what AI can safely do for regional care. The team proposes a framework to spur local data creation.
Themed contests invite providers continent-wide to build and share imaging sets. Curated, relevant data can grow fast under friendly rivalry. Prizes and visibility reward quality tagging and careful curation. Shared sets then train tools tuned to African health needs. Bias from foreign-only training data should shrink as local sets grow. The plan also builds skills inside the provider community itself. An inclusive AI ecosystem for the continent is the end goal.
Catalog record
| CREATOR | Mubaraq Yakubu, Udunna Anazodo, Maruf Adewole, Theodore Barfoot, Tiarna Lee, Tom Vercauteren, Jonathan Shapey, Andrew King, Alexander Hammers |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 21 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-00925
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Themed Challenges to Solve Data Scarcity in Africa: A Proposition for Increasing Local Data Collection and Integration Mubaraq Yakubu, Udunna Anazodo, Maruf Adewole, Theodore Barfoot, Tiarna Lee, Tom Vercauteren, Jonathan Shapey, Andrew King, Alexander Hammers Abstract In Africa, the scarcity of computational resources and datasets stand as significant hurdles impeding the development and deployment of artificial intelligence (AI) tools in clinical environments and contributes to the increase in global bias. These constraints complicate the realization of AI’s full potential, thereby posing significant obstacles to the advancement of healthcare in the region. This paper presents a proposed framework to address data scarcity in healthcare across the continent. The framework outlines a comprehensive strategy to encourage healthcare providers in Africa to create, curate, and share local medical imaging datasets. Through themed challenges, designed to promote participation, accurate and relevant curated datasets within the African healthcare community may be created. By facilitating this approach, the existing dataset scarcity may be overcome, fostering a more inclusive and impactful AI ecosystem tailored to African healthcare needs. 1 Introduction In the landscape of African healthcare, the dearth of computational infrastructure and representative datasets has posed formidable challenges for the development of artificial intelligence (AI) technologies. On the other hand, AI is transforming medical research and patient care through its diverse applications across various fields [ 10 ] . Emerging from high-income countries, there are large and carefully-curated datasets like the UK Biobank [ 18 ] , The Cancer Imaging Archive (TCIA) [ 8 ] and the Alzheimer’s Disease Neuroimaging Initiative (ADNI) [ 25 ] . However, Africa’s contribution to publicly available neuroimaging datasets remains disproportionately low compared to developed countries or even other middle-income countries [ 24 , 29 ] . Deficits in infrastructure, imaging facilities nearly exclusively operated by the private sector (clinics) rather than hospitals or universities, unreliable electricity supply, lack of centralised secure image data storage such as Picture Archiving and Communcation Systems (PACS), limited specialist availability, and a lack of knowledge and resources have compounded the obstacles hindering the growth and sustainability of healthcare initiatives across Africa [ 11 , 23 , 14 , 5 , 4 ] . Africa has the highest genetic diversity globally, encompassing more than 3,000 distinct ethnic groups [ 6 ] , so the absence of representative datasets for these ethnicities clearly limits the efficacy of AI tools in Africa, thwarting efforts to implement computer-assisted diagnosis and intervention effectively [ 12 , 14 ] . The data gap not only stymies the development of AI-driven healthcare solutions but also perpetuates disparities in healthcare outcomes, widening the chasm of global health inequities [ 5 ] . Racial bias has been shown even for seemingly innocuous tasks like brain segmentation [ 13 ] . Addressing this issue is important for extending the transformative potential of AI to African healthcare. Open challenges in medical image analysis have been instrumental in the advancement of machine learning algorithms [ 19 ] . Drawing from the success of similar challenges, here we propose adopting this approach to gather local datasets that represent the African continent as part of themed challenges for specific brain diseaese(s), incentivising contributors to submit the best-curated datasets possible. By making data accessibility for others a prerequisite for participation, we aim to create a virtuous cycle through which data availability can grow. This endeavor can pave the way for dedicated AI tools tailored to local contexts. Equipping clinicians with AI tools trained on local datasets and trusted data can accelerate the adoption of computer-assisted diagnosis and intervention [ 20 ] , driving advancements in African healthcare. 2 Current Approaches and Efforts in Africa Efforts to address healthcare challenges in Africa are already underway through a range of innovative initiatives and collaborative endeavors. We first review these initiatives before highlighting challenges and potential solutions in the African context. 2.1 Open Data Initiative The Open Data initiative [ 28 ] , which will be launched at the Medical Imaging Computing and Computer Assisted Intervention (MICCAI) conference 2024, aims to foster collaboration and innovation in medical imaging research by promoting the sharing of diverse and inclusive datasets. With a focus on underrepresented populations and diseases, particularly from the African continent, the initiative seeks to address disparities in healthcare and promote inclusivity in AI research. Central to this initiative is the establishment of the AFRICAI repository, which will host publicly available high-quality medical imaging datasets. By encouraging researchers and data custodians to contribute datasets to the repository and adhering to FAIR [ 20 ] data principles, the initiative aims to facilitate collaboration, reproducibility, and awareness of the importance of inclusivity in developing robust healthcare solutions. 2.2 African Neuroimaging Archive (AfNiA) AfNiA is an initiative aimed at democratizing access to medical imaging data in Africa [ 3 ] . By aggregating brain MRI scans and clinical attributes from multiple clinics across the continent, AfNiA facilitates the development and deployment of AI imaging innovations. Prioritizing ethical data use and governance, AfNiA ensures that AI solutions directly benefit the local communities that contribute data. 2.3 SPrint AI training for AfRican medical imaging Knowledge translation (SPARK) Academy SPARK Academy trains a new generation of African AI experts in medical imaging, bridging the gap between AI knowledge and clinical application ( [ 2 ] ). It is an initiative from the Consortium for Advancement of MRI Education and Research in Africa (CAMERA) which is focused on improving (or sometimes even establishing) MRI education in Africa. 2.4 Brain Tumor Segmentation (BraTS) Challenge 2023: Glioma Segmentation in Sub-Saharan Africa Patient Population (BraTS-Africa) The BraTS-Africa Challenge addresses the difficulties in diagnosing and treating gliomas in Sub-Saharan Africa (SSA) [ 1 ] . Gliomas, the most common type of primary brain tumors [ 27 ] , pose significant diagnostic and treatment challenges due to their aggressiveness and resistance to conventional therapy. Late presentation of the disease at advanced stages and the use of lower-quality MRI technology exacerbate diagnostic difficulties in SSA populations. The BraTS-Africa Challenge seeks to develop and evaluate computer-aided diagnostic (CAD) methods tailored to resource-limited settings, where the potential for transformative healthcare impact of AI is highest by mitigating the workforce limitations. 3 Difficulties Around AI Development in Healthcare in Africa and Potential Solutions 3.1 Resource Limitation Many healthcare institutions in Africa operate with constrained resources, facing limited access to advanced computational infrastructure like GPUs and high-speed internet, which hinders the implementation of AI solutions. This scarcity affects data processing capabilities and restricts the development of sophisticated AI models. Additionally, the socio-economic disparities in low- and middle-income countries (LMICs) result in insufficient investments in healthcare and research. The high number of low-income countries (LICs) and LMICs in Africa means government funding for AI health research is often inadequate [ 21 ] . This lack of investment hinders the acquisition of essential clinical data-generating devices, such as imaging equipment. Even when these devices are obtained, maintaining them—like MRI scanners—becomes challenging due to frequent breakdowns caused by intermittent electricity supply [ 4 ] . Addressing these resource limitations requires increased funding and investment in basic infrastructure, healthcare, and research infrastructure. International collaborations, public-private partnerships, and initiatives aimed at enhancing technological capabilities in LMICs can play a pivotal role in overcoming these challenges. Additionally,