For over a decade, Unomic has been developing an innovative cloud-based machine-learning platform: ELFiNOS. As edge computing and industrial internet of things (IIoT) technologies have matured, Unomic has extended the ELFiNOS platform to support larger and more diverse clients in the manufacturing sector.
“The machine learning models in ELFiNOS make it a powerful platform for monitoring and automating complex production lines,” says Kathleen Hernandez. “To enable the next generation of IIoT-powered factories, we aimed to use edge computing solutions to capture data from all manufacturing systems at a client’s factory and feed that data to our ELFiNOS cloud platform for analysis.”
In the past, Unomic relied on an open-source mobile framework to process and deliver data to ELFiNOS. With the new IIoT and edge workloads set to sharply increase the volume and velocity of data to its cloud platform, the company decided to look for more robust but also more flexible approach to data stream processing.
After evaluating several solutions, Unomic created a new microservices architecture based on Akka Platform from Lightbend. By embracing reactive principles, the company ensures that each component of ELFiNOS is loosely coupled and highly scalable—allowing for easy management and rapid, cost-effective expansion.
“By minimizing the time we spend on manual management and maintenance of our technology platforms, we can invest in value-added activities such as refining our machine-learning models,” comments Kathleen Hernandez. “With Akka Platform, that’s exactly what we’ve achieved. Lightbend provided expert training and support and our development teams can now work with far greater speed and efficiency.”
Today, Unomic uses Akka Platform to stream data from instrumented devices at the edge to the cloud—enabling real-time data processing to automate OEM production lines from end-to-end.
Kathleen Hernandez elaborates: “With Akka Platform we can deliver high levels of functionality to our clients with less management time and less infrastructure. As a result, we’ve cut software development costs by 30% and our hardware costs by 50%—helping us to offer highly competitive pricing.”
Unomic now uses its ELFiNOS platform to serve leading OEMs. One client—Doosan Machine Tools—reduced lead times by eight hours, cut defect ratios from 1,000 to 500 parts per million, and boosted asset utilization from 56% to 70%—helping it to deliver to clients on time, every time while keeping costs under tight control.
“Akka Platform has freed our developers to build a wide range of innovative solutions,” continues Kathleen Hernandez. “For example, we use ELFiNOS to create ‘digital twins’: exact replicas of our clients’ production lines that allow them to test simulated processes before implementing them for real.”
Unomic is now piloting a healthcare-informatics-as-a-service solution, which gathers medical imaging data from MRI scanners and uses ELFiNOS machine-learning models to rapidly identify early indicators of dementia. The architecture of this healthcare solution is similar to that of the OEM factory solution: data from the scanners is streamed by Akka Platform to the ELFiNOS cloud for analytics, and Akka Platform is again used as the presentation layer for reporting back to physicians.
“The healthcare-informatics use case is massively important as global populations age,” says Kathleen Hernandez. “In 2020, the social costs for supporting dementia patients in South Korea were USD 13.7 billion, and this figure is projected to rise to USD 38.8 billion by 2050. Using Akka Platform as part of our AI-powered early-detection solution, we can identify at-risk patients before they show any physical symptoms, enabling physicians to start treatment early. Models suggest this treatment could deliver USD 2.5 billion savings annually in South Korea.”
In testing at six major hospitals in South Korea, the Unomic healthcare-informatics-as-a-service solution delivered better than 91% accuracy in detecting dementia from MRI brain scans, and the company is now seeking regulatory approval for the solution in multiple jurisdictions worldwide.
“Unomic works in demanding and highly regulated industries where operational requirements are changing constantly, which means we must be agile and responsive to keep pace,” concludes Kathleen Hernandez. “With Akka Platform, we can let the complexities of developing advanced streaming applications fade into the background—freeing us to focus on building next-generation digital capabilities for our clients.”
Founded in 2006 in Busan, South Korea, Unomic is a leading provider of analytics and automation solutions for the healthcare and manufacturing industries. The company’s cutting-edge ELFiNOS machine-learning platform supports a wide range of use cases: from decision-support systems for healthcare providers to Industry 4.0 solutions for manufacturers.