FedML, a decentralized and collaborative machine learning platform that enables training, deployment, and continual improvement of AI models anywhere, announced today that it would partner with Konica Minolta, a Japanese multinational technology company manufacturing business, industrial and medical imaging products, including copiers, laser printers, multi-functional peripherals (MFPs), digital print systems for the production printing, textile printers, medical and graphic imaging products, such as X-ray image processing systems, ultrasound systems, color proofing systems, and X-ray film; photometers, 3-D digitizers, and other sensing products.
“Large machine learning (ML) models are breaking new ground every day, achieving unprecedented performance in various application domains, such as computer vision, speech and language processing, data mining, healthcare, and life sciences,” said Salman Avestimehr, CEO and Co-Founder of FedML. “However, the performance of such models heavily depends on the size and diversity of the datasets, as well as the quality of their annotations. This creates a major pain point in the development and utilization of state-of-the-art ML models in the healthcare domain, as medical data obeys strict regulatory laws and privacy restrictions and is very costly to annotate. This leaves most institutions with their own datasets, limiting AI from reaching its full potential in health tech.”
Also Read: Publicis Health Launches Insagic, a Groundbreaking Insights and Advisory Business
Fortunately, FedML allows AI technologies to use these siloed datasets without centralizing or transferring the data. It enables machine learning from decentralized data at various nodes without concentrating any data in the cloud , which provides privacy, reduces development costs, and furthermore empowers data owners to monetize their data and compute resources.
“It has become clear that machine learning and AI can greatly impact healthcare by improving diagnostics, patient safety, and treatment planning. According to the recent market research data, the projected market size of AI-based global healthcare solutions is expected to reach $200B+ by 2030. However, healthcare data, which is essentially the fuel for machine learning, is very much siloed and heterogeneous due to privacy rules (e.g., HIPAA, GDPR, CCPA, etc), data regulations, disconnected IT infrastructure of healthcare systems, and diverse procedures for data collection and annotation at different healthcare institutions.
SOURCE: Businesswire