Artificial intelligence (AI) will become one of the most important factors affecting the development of human society in the coming years. We put into this concept all areas of development of the sphere, including Machine Learning (ML), Generative Adversarial Networks (GAN), Gradient-boosted-tree models (GBM), deep reinforcement learning ( Deep Reinforcement Learning, DRL), etc.
Acting as a cloud provider, Cloud4Y has partnered with various healthcare organizations. It was always an interesting experience with new technical, legal, psychological difficulties that had to be overcome.
Business, technology, and healthcare are areas where AI is most in demand. Let's take a look at how AI / ML tools can influence the quality of healthcare delivery.
The idea of using artificial intelligence in medicine dates back to 1972, when MYCIN was launched at Stanford University. It was a prototype AI program used to study the issue of blood poisoning. Early AI research continued mainly in American institutions (MIT-Tufts worked together, actively developed the technology at Stanford and Rutgers University. In the 1980s, Stanford University continued its work in the field of artificial intelligence as part of the project "Medical Experimental Computer Artificial Intelligence in medicine ”(SUMEX-AIM).
Thanks to the growth in computing power and the emergence of new artificial intelligence technologies, work in this direction has become much more active. News regularly appears about the next scientific discovery made with the help of neural networks and machine learning. What interesting things can you tell about the possibilities and prospects of AI in medicine today?
AI in radiology computers questions
Numerous medical imaging data is stored in abundance in small local systems. But what if you leverage deep learning by uploading data to the cloud and feeding it to AI? Machines and algorithms can efficiently interpret imaging data by identifying patterns and anomalies.
Most obvious use case: a radiologist / radiologist assistant, who is involved in identifying and localizing suspicious skin lesions, lesions, tumors, internal hemorrhages, brain lesions, etc. The computer works faster and more accurately, and therefore is able to give out specific data about the disease a few seconds after processing the information. Man cannot do that.
There is another point. Highly qualified specialists are expensive and in great demand. They are under serious pressure, literally bogged down in the streams of data that are pouring on them from all sides. If you believe this article , such a specialist should issue a diagnosis every 3-4 seconds. Machine intelligence can improve the skills of the ordinary specialist, helping him to sort out difficult situations. Thus, reducing the number of false diagnoses and saving lives.
The identification of rare or difficult to diagnose diseases often depends on the experience of the doctor, as well as the degree of "neglect" of the disease. Simply put, until the sore gets out, it may not be recognized. By training a computer on large datasets containing raw images and many forms of pathologies associated with certain diseases, it is possible to improve the quality of diagnosis and the number of diseases identified. This idea is being developed by the startup AIDOC.
AI is able to improve the quality of the work of medical institutions by automating the time-consuming and responsible part of the work of doctors. With the help of computer algorithms, you can also control the effectiveness of treatment and the quality of the operation performed, and predict the rate of recovery of the body.
Microsoft's InnerEye project is a good example of such technology. He suggests using ML techniques to segment and identify tumors using 3D X-rays. This can aid in accurate surgery planning, navigation and effective tumor contouring for radiation therapy planning.
AI in pathology
Pathological diagnosis includes examination of a tissue section under a microscope. Using Deep Learning to train an image recognition algorithm, combined with human experience, will provide more accurate diagnostics. Analyzing digital images at the pixel level can help detect lesions that the human eye can easily miss. And this will provide a more efficient diagnosis.
Such technology is being developed, for example, by Harvard Medical School. The algorithm uses speech and image recognition technology to recognize images with pathologies and trains computers to distinguish between cancers and non-cancers. Combining this algorithm with human work resulted in 99.5% accuracy.
Machine Learning and Medical Science
Petabytes of data are generated in all kinds of medical facilities. This data, unfortunately, is usually randomly scattered and unstructured. This is by no means a reproach towards doctors. They have not so much to treat as to report on treatment. However, chaos greatly interferes with planning and global monitoring of the health of a particular country or the world as a whole.
An added complication is that, unlike standard business data, patient data does not lend itself well to simple statistical modeling and analytics. A powerful AI-powered cloud platform with access to medical databases is capable of efficiently analyzing mixed information (eg, blood pathology, genetic traits, X-rays, medical history). It is also (theoretically) capable of analyzing input data and revealing hidden patterns that are not visible due to an excessively large amount of medical information.
Interpretable AI models and distributed machine learning systems are great for these tasks. They will allow not only to effectively develop medical science, finding new patterns and racial / sex / age characteristics of people, but also to form more accurate data on the health status of the population in specific regions.
Surgical assistant robots
Already, many operations are carried out using computer vision and manipulators controlled by a surgeon. This is a significant part of the development of medical technologies, leveling the factor of human fatigue and increasing the efficiency of procedures. AI robots are great at helping conventional surgeons. For example:
Supervise the work of the doctor, acting as an insurance against inattention;
Improve visibility for the surgeon, remind him of the sequence of actions during the procedure;
Create accurate, minimally invasive tissue incisions;
Reducing the pain level for the patient through the selection of the optimal incision geometry and suture.
No comments:
Post a Comment