Auriga, a custom software development service provider, recently participated in the Artificial Intelligence (AI) Forum. Auriga joined the event to expand its cooperation with French companies and discuss AI trends with business leaders, researchers, and authorities.
The AI Forum covered several AI application areas, including healthcare, manufacturing, environmental protection, urban planning, education, scientific research, and public administration. Auriga’s experts participated in a session devoted to AI technologies in healthcare, discussing the challenges and barriers to AI adoption in hospitals and clinics and how to overcome them with reasonable solutions.
AI vs. Hybrid Intelligence
In recent years, AI has been hyped as a game changer in many industries, including healthcare. At the same time, the revolutionary—and scary—thing is that AI is not an extension of humans but a substitution. AI does not help in analysis and decision-making; it makes its own decisions. However, it is still hard to imagine that healthcare patients will want to fully entrust their lives and well-being to AI algorithms.
At the AI Forum, all speakers of the “AI in Healthcare” session agreed that AI is not the correct term to describe the solutions healthcare professionals, clinicians, and patients require today. It is rather hybrid intelligence (HI), which combines human and machine intelligence. Medical HI is intended not to replace physicians but to help them make better decisions faster and accelerate diagnostics by quickly analyzing digital patient data.
Playing Big with “Small Data”
HI-enabled solutions rely on AI subfields, including machine learning and computer vision. No one would underestimate these technologies’ potential, especially in healthcare, where they promise such benefits as treatment personalization and enhanced medical procedure efficiency. However, machine learning and computer vision heavily depend on input data, and healthcare is known for its challenges in obtaining this data, including sufficiency, reliability, and security.
Auriga’s report at the AI Forum showcased our experience efficiently training neural networks for medical tasks when accessible patient datasets are relatively small. Auriga’s speaker, Andrey Teterin, shared our ECG rhythm classification takeaways using insufficient labelled data for cardiac monitoring system development.
Summing Up
Andrey Shastin, Auriga’s Head of Global Business Development in the medical and embedded systems domains, said the following:
The AI Forum was a suitable platform for us to demonstrate how Auriga could help medical technology companies cope with some of the challenges of adopting emerging digital technologies in healthcare. In particular, we presented a solution to increase the reliability and efficiency of neural networks when labelled data is insufficient. Our experience shows that the healthcare community is more enthusiastic about machine learning and computer vision technologies than full-scale AI adoption.
Vyacheslav Vanyulin, Auriga’s CEO, added this:
Being a software development company experienced in not only the medical domain but various domains, Auriga offers the concept of seamless addition of AI-related methods and technologies to medical solutions of different classes. Full-scale AI adoption is still a thing of the future in such a delicate industry as healthcare.
Auriga boasts 18 years of medical software development and testing experience, and both customers and industry experts, such as the IAOP®, have repeatedly acknowledged it. The company is ISO 13485 certified and runs eight R&D and testing labs for medical devices in Eastern Europe. Explore our online portfolio of medical device and digital health projects, or contact us via the website for more information on Auriga’s services.
