minoHealth: Ghanaian Team Develops Artificial Intelligence Healthcare Systems That Outperform Human Doctors
minoHealth is a health system developed by a Ghanaian team. The technology based health system uses Deep Learning (Artificial Neural Networks) in Artificial Intelligence (AI) to predict and diagnose medical conditions in patients.
These AI systems are said to outperform human physicians in diverse ways. minoHealth currently has three AI healthcare systems.
- The first system predicts if a female patient would develop Diabetes in the next 5 years or not.
- The second and third systems determine if a Breast Tumor is Malignant or Benign with two separate approaches.
Deep Learning is the most effective part of Artificial Intelligence today, it’s in fact used by various top tech firms today to solve various kinds of problem, like Google DeepMind for AlphaGo which beat human Go players, Facebook for AI algorithms that sort posts on your newsfeed and auto tag your images, and Tesla Motors for self-driving cars.
Deep Learning is developed loosely based on the human biological neural network, so rather than give it direct instructions, AI scientists and engineers train these Deep Learning models with datasets and the models automatically learn from them like a child learns by observing their environment.
So minoHealth applies Deep Learning algorithms to Healthcare problems in order to develop AI systems that better diagnose and predict medical conditions than humans. The team also plans to work with Epidemiologists in Ghana and Ministry of Health to develop lots of medical datasets to train other Deep Learning models to cater to even more medical conditions and healthcare needs of Ghanaians.
The team is also currently in talks with Ghanaian Hospitals so they can democratize their AI systems to be used for all Ghanaians.