Leifeng.com, recently, Aster Innovation Research Center, a subsidiary of medical group Aster DM Healthcare, cooperated with Intel Corporation and artificial intelligence company CARPL.ai to develop and launch an AI-based health data platform in India.
Headquartered in the United Arab Emirates, Aster DM Healthcare is one of the fastest growing healthcare groups in the MENA region with more than 20,000 employees worldwide. At the same time, the company also has a presence in the Indian medical field.
It is understood that this health data platform is based on federated learning technology, which can train AI algorithms across multiple decentralized data sources (including local data samples) without data exchange. In addition, the platform also applies OpenFL, Intel’s open source framework for training machine learning algorithms, to facilitate the development of federated learning algorithms. OpenFL can be combined with CARPL.ai’s data extraction, transformation and loading capabilities for end-to-end AI model training.
It is understood that the health data platform has been tested using hospital data from Aster Hospital in Kerala, Bangalore and Vijayawada, and more than 125,000 chest X-ray images have been extracted, using the two-site method to train CheXNet AI model and detect anomalies in x-ray reports by the model.
Aster DM Healthcare officials said that the advantageous role of AI solutions in medical imaging is obvious to all, and it is more helpful to solve urgent problems such as shortage of medical staff, but how to access data silos in medical institutions and other medical systems while complying with regulatory policies Still a “great challenge”.
According to Nivruti Rai, head of Intel India, the key imperatives for developing AI applications are obtaining high-quality training datasets and addressing limitations in the form of regulatory frameworks and geographic boundaries.
By providing access to huge datasets, this federated learning-based platform developed by Aster DM Healthcare enables teamwork to develop AI-enabled health technology solutions and further advance drug discovery, diagnostics, genomics and healthcare prediction innovations in other fields.
At the same time, the platform also allows clinical trials to access relevant datasets in a secure and distributed manner.
As a service provider, the platform is expected to improve the accuracy of AI model training, while enabling data scientists from different organizations to perform AI training without sharing raw data. With security and privacy guarantees, the platform also ensures compliance and governance of organizational data.
According to Aster DM Healthcare, its recent pilot also demonstrated how the platform could enable “democratic access to health data across organizational and geographic boundaries without compromising data privacy and security.”
In recent years, Aster DM Healthcare has made great strides in expanding the application of AI technology in the Indian healthcare sector.
One of the important performances is the AI laboratory opened by Aster CMI Hospital, a general hospital in Bangalore. In March this year, Aster CMI Hospital and the Indian Institute of Science jointly opened the Aster AI Lab, which aims to build AI healthcare tools and train AI healthcare professionals.
In its initial stages, the lab will first focus on developing AI tools for neurology, and then expand to other clinical specialties.
Rai declared that the development of a federated learning-based health data platform “marks a paradigm shift in ‘translating computation into data’ rather than ‘transforming data into computation'”.
Dr. Azad Moopen, chairman and founder of Aster DM Healthcare, said: “So far, only a few such measures have been carried out in the healthcare sector.” Their health data platform will “support the development of predictive mechanisms for patients, providing first-hand information on treatments. Opportunities for second opinion and, most importantly, confirmation of patient data security and confidentiality.”
Dr. Vidur Mahajan, CEO of CARPL.ai, pointed out: “There is no doubt that decentralized data storage and subsequent joint training of AI models are the future direction, especially in the context of AI’s lack of generality.”
Source: https://ift.tt/3VxeHb7
This article is reprinted from: https://www.leiphone.com/category/healthai/xZjxOqyZmJJb4jvZ.html
This site is for inclusion only, and the copyright belongs to the original author.