AI-enabled solutions in areas such as medical imaging help address pressing challenges in healthcare such as staff shortages and the elderly population. However, accessing the silos of relevant data spread across different hospitals, geographical areas and other health systems is a major challenge when adhering to regulatory policies.
Commenting on the collaboration, Dr. Azad Mupen, Founder, Chairman and Managing Director Aster DM Healthcare said, “The Secure Federated Learning Initiative will help analyze data and develop a predictive approach for patients, providing a second opinion on treatment and, most importantly, supporting patient data security and confidentiality. This collaborative platform with world leaders will open the door for many players in the field to participate in developing accessible healthcare solutions.
“AI applications are at the forefront of revolutionizing healthcare through timely and effective screening, diagnosis and treatment of diseases,” said Nivruti Rai, Country Head, Intel India and Vice President, Intel Foundry Services. This solution will be offered by both AI researchers and data custodians as a service used in the pursuit of advancing AI innovation and wider impact in healthcare. It marks the paradigm shift by getting the computer into the data instead of the computer getting the data. ”
CARPL.ai’s Chief Executive Officer Dr. “There is no doubt that federation of de-centralized data storage and subsequent training of AI models is the future, especially as the lack of generalization of AI is becoming a major problem. Our mission is to take AI from the bench to the clinic,” he said. Is to go, and this is another example. “
Federated Learning (FL) is a method of training AI algorithms with data stored on multiple decentralized sources without moving data. To facilitate the adoption of federated learning, Intel has led the development of an openflopen source framework to train machine learning algorithms that take advantage of Intel’s security technology to provide a solution for “data silos”.
The capability of this platform was demonstrated using hospital data from Aster Hospital’s Kerala, Bengaluru and Vijayawada clusters. More than 125,000 chest X-ray images, including 18573 images selected from more than 30,000 unique patient data from Bengaluru, were used to train the CheXNet AI model using a two-node / site approach – in Bengaluru and Vijayawada – X Federated Learning -Ray report to detect abnormalities. 18,573 unique images, in addition, provide a 3 percent accuracy boost due to real-world data that would not otherwise be available to train the AI model.
The success of this pilot has shown the connection to the next level, which is the democratization of access. Health data Cross organizational and geographical boundaries without compromising with Data privacy And security aspects.
