How machine learning can help us in medicine

Hello everyone and welcome to yet another promising day. Today I want to talk to us about machine learning and how it can help us in the field of medicine.

^{Image source Maxpixel. CC0 creative commons license.}

One would ask, what is machine learning? We all are intelligent in our own different ways, most of the time, our intelligence depends on our exposures in life, hence, it is very difficult or almost impossible to figure out a problem that you've not come across before or come across its kind. We are more intelligent now than when we were born, this is because "programs" in our head is made to "learn" as we are exposed to different things in life. The same thing is applicable in machine learn, machine learning is a part of artificial intelligence which deals with improving the ability of an already existing system without the need to manually impact such abilities.

A typical example of machine learning is the fingerprint system in our mobile phones or the retina scan security device on most of the highly secured facilities. These systems, for example, the fingerprint scanner in our device by default cannot recognize our fingerprints but when we expose our fingerprints to it couple of times, it becomes more intelligent and will be able to recognize it in subsequent times and even use it to do more amazing things.

Machine learning in Medicine

Doctors on daily basis make calculations, extensive reasoning, diagnosis, reference different materials and also perform lots and lots record keeping which is later used as reference points and conclusions can be drawn based on the references to these resources. Once a computer learns how to perform a task, whether the learning was automatic or the knowledge was manually impacted on the computer, to a great degree, it can perform these functions better than humans.

^{Some of the basic machine learning process. Image source Wikimedia. CCv3.0 licence}

Some of these processes are routinely carried out and the analysis of activities like test results and correlations can be assigned to machines or robots. It is no news that some voice assistant present in some hospitals can predict to a great extent, someone’s illness just by listening to them speak about their symptoms. This action is only possible because the voice system makes reference to a very large data set which it references with a very great speed to come up with result.

Hence, the idea of big data is also leveraged in machine learning applied in medicine. Big data has been a computing field with a very rapid evolution and it involves a complex analysis of data which vary widely and to ordinary human brain, such details are overwhelming. Hence, the concept of big data takes data such as comments on Facebook from millions or billions of users about a particular product and create information about the future of such product.

Machine learning has been applied in the field of medicine in the following ways:

Diagnosis

When is say that machine learning can be applied in the field of diagnosis, this does not rule out the function of physicians, rather machine learning only aides their duties. As already stated, we are as intelligent as our exposures in life and the same can be said for virtually anyone. That a physician is good is mostly due to his experience, either in theoretical world or in practical world. Hence, a physician who has been exposed to issues relating to stomach for a long period of time will be more poised to solve stomach related issues than a physician that has been exposed to issues relating to eyes or bones.

^{A special voice assistance system that can setup medical records and also assist in analyzing patients based on their stated symptoms. Image source US Air Force}

Due to the limitations we have as human, we created the concept of specialization in order to improve productivity but the same cannot be said for machines. Depending on the capacity of an intelligent system, it can be equipped with vast information about different health issues and their possible solution and also with the help of the human it is currently working with, it can learn and further improve on its database and as such can make very quick and valid health recommendations and carry out risk analysis of patients.

Perception duties

By examining the X-ray scan of a patient, a doctor can state with a high degree of accuracy, the skeletal condition of his patient. Radiologists constantly examines images and scan results in order to diagnose a patient. Intelligent machines and robots has been designed using complex algorithms to be able to examine radiological images and identify medical abnormalities based on reference to a normal condition.

^{Google's algorithm can help detect breast cancer from examining mammograms. Image source Military health}

In 2017, Google recorded a success in development of some algorithm which can be used in the healthcare for identifying cancerous growths in human mammograms. The process involves the rigorous examinations of mammograms with high speed and accuracy which beats the human performance. Also the Stanford University has recorded success in its move in the development of algorithms that can identify skin cancer in its earliest stage.

Conclusion

Some medical practitioners believe that the involvement of AI and artificial learning could render them jobless, but this is untrue as quick recovery of patients will always involve human intervention and human care.

By analyzing large amount of preexisting medical records and information, intelligent systems can be a better assistant to medical practitioners as stated above.