2022 | Early stage | Health | SDG11 | SDG3 | United States
AutoHealth: Automatic Deep Learning Software for Low Cost Healthcare Assistance

Company or Institution

A collaboration between IBM Research, Al Akhawayn University, and UPHF France






United States

Sustainable Development Goals (SDGs)

SDG 3: Good Health and Well-being

SDG 11: Sustainable Cities and Communities

General description of the AI solution

Artificial intelligence improves the lives of patients, physicians, and hospital managers by assisting in the detection and segmentation tasks in a fraction of the time and at a fraction of the cost. While, the AI healthcare sector has benefited with up to 20% of the overall AI fundings, its applications and research are still far from applicable in poor countries. Indeed, Africa's healthcare systems suffer from neglect and under-funding, which prevent the use of AI and specifically deep learning algorithms that require computationally intensive resources.

In north Africa, the most common causes of death are heart diseases and cancer. With Arterial febrillation and hypertension at the top of the stack. Early the detection can be critical in saving lives and giving patients a second chance of battling these diseases. However, due to the restricted and poor resources, many African regions and hospitals do not offer this critical service. Hence, the disease is caught at the last stage and after causing severe problems.

We propose a platform and solution that aims at assisting radiologists and physicians in their daily diagnositcs by finding the most-efficient and accurate algorithm for a specific application. The algorithm should be deployed on a tiny microcontroller and embedded devices that are extremely cheap. Depending on the situation and use case, our software can offer the following capabilities:
– Reorder the stack of screenings a radiologist need to analyze to make sure critical cases are analyzed first.
– Offer a second eye diagnosis that can guide the physician throughout the different patients' encounters.
– Analyze motion compensated pulse rate and detect arterial fibrillation and hypertension from a wearable device.


– Hadjer Benmeziane, Hamza Ouarnoughi, Smail Niar, Kaoutar El Maghraoui, CaW-NAS: Compression Aware Neural Architecture Search, 25th Euromicro Conference on Digital System Design (DSD), 2022
– Hadjer Benmeziane, Smail Niar, Hamza Ouarnoughi, Kaoutar El Maghraoui, Pareto Rank Surrogate Model for Hardware-aware Neural Architecture Search, IEEE ISPASS 2022
– Hadjer Benmeziane, Hamza Ouarnoughi, Kaoutar El Maghraoui, Smaïl Niar: \textit{Real-time style transfer with efficient vision transformers}. EdgeSys@EuroSys 2022: 31-36
– Hadjer Benmeziane, Hamza Ouarnoughi, Kaoutar El Maghraoui, Smail Niar. Accelerating Neural Architecture Search with Rank-Preserving Surrogate Models. 7th International Conference on Arab Women in Computing (ArabWIC 21). \textbf{Best paper award}.
– Hadjer Benmeziane, Kaoutar El Maghraoui, Hamza Ouarnoughi, Smail Niar, Martin Wistuba, Naigang Wang. Hardware-Aware Neural Architecture Search: Survey and Taxonomy. 30th International Joint Conference on Artificial Intelligence (IJCAI-21)





Public Exposure

Mentorship Program

HPC resources and/or Cloud Computing Services


International Research Centre
on Artificial Intelligence (IRCAI)
under the auspices of UNESCO 

Jožef Stefan Institute
Jamova cesta 39
SI-1000 Ljubljana



The designations employed and the presentation of material throughout this website do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any country, territory, city or area of its authorities, or concerning the delimitation of its frontiers or boundaries.

Design by Ana Fabjan