Jan 21, 2021

Data For Good Fellowship Recipient Wants to Empower Rural Healthcare Providers with AI

Mahdieh Taher, a Fall 2020 Master of Data Science graduate, has been awarded the Jack Larson Data for Good Fellowship for her work as a data engineer dedicated to improving medical diagnoses in rural areas. She believes that data science can help tackle some of the toughest challenges facing society today, and her ultimate goal is to leverage artificial intelligence in medicine to improve the lives of those in rural communities.

Taher was in the data science class Deep Learning in the Cloud and at the Edge (W251) when she had an idea: doing machine learning and deep learning on edge devices — computers and other devices at the “edge” of the network, like a rural doctor’s laptop or smartphone that she brings from location to location — could bring much-needed medical diagnostic capacity to remote populations. If the computation can be done on edge devices themselves instead of on large, powerful, centralized servers, it removes the need to transmit large data files using the often unreliable network infrastructure available in remote locations; solving the problem of poor internet connectivity preventing the use of effective cloud-based diagnostic solutions in rural areas or other isolated communities.

Taher envisions designing an edge device that doesn’t require robust internet connectivity to detect diabetic retinopathy.

Diabetic retinopathy — one of the main causes of vision loss in the world — is of particular interest to Taher, whose grandmother lost her vision to the disease. “Because diabetic retinopathy doesn’t have noticeable symptoms in early stages,” Taher explained, “doctors recommend all diabetics have an annual eye exam. The problem is that in many rural areas, ophthalmologists are in short supply.” This means the disease might go undetected for years, resulting in preventable vision loss.  

Taher envisions designing an edge device that doesn’t require robust internet connectivity to detect diabetic retinopathy. Such an AI-driven device would enable many different healthcare workers to be trained to use it to compensate for the lack of ophthalmologists in many rural areas, leading to efficiency and increased quality of care for diabetic patients. Taher is currently collaborating with a small team to make her idea a reality, including a local machine learning engineer and a research scientist based in Paris. 

The Jack Larson Data for Good fellowship supports Master of Information and Data Science students who desire to use data science to benefit society. Successful applicants have a background working for non-profits, government, or community service organizations or be able to demonstrate how they intend to use the data science skills acquired in the MIDS program for the betterment of society.

Last updated:

January 21, 2021