Do health professionals need to learn code?
Why so much drama with Ai in Health?
It is undeniable that Artificial intelligence is gaining the most significant attention in every aspect of human life. From the most common activities like selecting your best picture of the month or recommending new series on streaming platforms to situations that seem out of comic books like self-driving cars and chatbots sounding exactly like humans, AI is merging so smoothly within our life that we barely realize it is here. And of course, the healthcare industry is not the exception. The performance of AI in healthcare is so groundbreaking that some algorithms, like the DLAD developed by Seoul National University Hospital, are even outperforming specialists worldwide.
However, if we are fair, the healthcare industry is not only “not the exception” but is perhaps one of the most significant aspects that AI is revolutionizing. In simple terms, AI is helping doctors, hospitals, insurance companies, and the pharmaceutical industry to make better decisions for the well-being of patients. All these applications have only been possible thanks to the rapid digitization of the healthcare ecosystem. In fact, in the last report of the International Data Corporation (IDC), it is estimated that medical data would grow even faster than data from the financial services or manufacturing industries. With this growth, the amount of data available for AI experts to take advantage of has been unthinkable, making the healthcare industry the ideal setting for applying technological tools.
But developing an Artificial Intelligence algorithm (until now) is not an easy task. Cloud Moyo, a Microsoft partner for applying AI and cloud computing in companies, has identified 10 problems when working with AI. Of these, I would like to highlight three that are of great relevance to healthcare: “Lack of technical knowledge,” “data acquisition and storage,” and “Rare and expensive workforce.” Now, suppose we leave aside the difficulties of collecting, storing, and labeling the data. In that case, there still is a massive barrier for clinicians and health professionals to work with AI: WRITING CODE.
The problem of writing code is quite significant. From our experience, doctors, nurses, bacteriologists, or any healthcare professional qualified to write codel correctly are quite a few. Even more, the ones prepared to write an AI algorithm following the highest standards can be counted on one hand. Of course, hiring a whole tech team to take care of the developing part is always an option, but it is an expensive and time-consuming process since the technical abilities required for developing AI algorithms are relatively rare in the current market place
So, Healthcare professionals should learn how to program to develop AI solutions?
The short answer is: Definitely not. Although several undergraduate programs are highlighting the importance of programming in the modern world, there are so many skills that healthcare professionals must develop to carry out all the tasks in their daily job properly that adding programming to them is almost disrespectful. For years healthcare staff has understood the vital role of data in the evolution of medicine. Therefore, many of them have started to collect relevant data for specific conditions, anticipating the day when AI becomes so easy to use that they become the developers of the algorithms. Actually, in Arkangel, we believe that only when healthcare professionals become designers and developers will we be able to see the real impact of AI on patients’ health.
That is why in the last years, we have been working on the development of a platform that allows physicians to create their models without the need of knowing how to write code.
What if I tell you that you do not need to write a single line of code to train, validate and test an Artificial intelligence algorithm?
Well, that is precisely what Hippocrates, the ArkangelAI technology team’s latest developments, is designed to do. Hippocrates is a technological platform that receives healthcare data and automatically trains, validates, and tests an AI algorithm for a specific medical task. The whole platform is built to work precisely how a group of AI expert engineers would approach developing a new algorithm, but without the need for the user to write a single line of code. That is to say:
- Hippocrates helps to structure the data.
- Hippocrates receives the previously prepared data.
- Hippocrates splits the data into the well-known train, validation, and test sets.
- Hippocrates adjusts the model architecture so that it best fits the specific task required by the user.
- Hippocrates trains the model in such a way that it maximizes the performance metrics
- Hippocrates returns the model to the user showing the obtained metrics.
- The users can distribute their models on Arkangel’s platform or/and on their local infrastructure.
This entire process can take place on-premise or in Arkangel’s cloud service. “We have been using Hippocrates to develop some of our algorithms and evaluating its performance with some allies, showing quite astonishing results. Jose Zea, CEO of Arkangel Ai”
Do you want to know more about Arkangel Ai?
Do you want to know more about how Hippocrates AutoML works?
Contact us without hesitation, surely you have in mind an important and relevant innovation for humanity, if so Hippocrates will be in your hands and your ideas will become your legacy.