Machine learning (ML), sometimes referred to as Artificial Intelligence (AI), might sound complicated and seem like a technology of the future to a non-tech person. However, already today the vast majority of us interact with ML and AI technologies at some level without realizing it. In this article, we are going to take a look at how each of us interacts with machine learning and artificial intelligence on an everyday basis.
Simply speaking, machine learning helps humanity analyze big sets of unsorted data, draw patterns, and find dependencies to drive analysis and discover tendencies that otherwise could have gone unnoticed. ML and AI became irreplaceable, namely in healthcare and social media. What could be some of the real-life examples of such applications?
In a connected world where every device can potentially generate, store, and transfer information, it became possible to collect and compare all sorts of data through crowdsourcing. For example, X-ray images from numerous research centres and medical institutions, heart rates across different groups of the population through health trackers, location and movements of people and objects through GPS systems in mobile devices, preferences in food, literature, gaming through search history, etc.
As regular citizens, we experience the impact of machine learning every time a social network suggests a new page or user to follow based on what you have liked or searched before on the platforms, or a medical device provides test results with a 99,8% accuracy relying on results obtained from tens of thousands of similar cases.
Such services and high-levels of precision are only possible because crowdsourced data sets are broader than any set that an individual doctor or an industry expert could have encountered in one’s life. However, this is not the limit to what machine learning could do in the future. As a constantly learning and evolving technology, it provides a broader spectrum of applications.
The aim of precision medicine is to provide patients with treatment recommendations based specifically on their health history and peculiarities of their condition, including genetics. One more area of application is to create systems that would set up a patient’s dose of medicine by tracking information on their blood, nutrition, sleep, and the stress level, etc. In order to make sure you will not forget to take the medicine in the correct amounts.
“Imagine if you could take results of all of the tests … and the results of the treatment that was done, and aggregate and anonymize all of that data, and apply machine learning to learn from that which treatments were the most effective. Not only could you reduce the amount of the chemotherapy that was required for a patient, but you could also reduce the number of patients who received an unnecessary dose – or who received a type of chemotherapy that didn’t work.”Mike Flannagan, senior vice president of SAP Analytics.
Monitoring and prevention
The more data machine learning systems collect, the better they become at discovering trends and co-dependencies in data sets. For example, Instagram algorithms can detect a decline in a user’s mental health state by observing posts and search history and offer help by sending notifications and messages with contact information of crisis centers and counseling lines.
In a similar way, medical wearables can detect changes in and potential threats to one’s health by monitoring such vital information as heart rate, blood pressure, sugar and insulin levels, etc. and inform both the patient and a doctor about findings. Shifting from treatment of acute states and late-state diseases to early detection and preventive treatment will directly impact the quality of life and longevity of the humankind.
One of the most famous examples is the da Vinci robotic surgical system that enables surgeons to manage deft robotic arms and carry out the least traumatic ad very detailed surgical manipulations that are impossible for humans to perform. One day, machine learning might be utilized to bring together visual data and motor patterns by means of devices like the da Vinci to enable machines to perform surgeries.
There are some systems employing computer vision to detect distances and body parts. Moreover, ML is sometimes utilized to stabilize the motion and movement of robotic arms manipulated by human professionals. This kind of machine might be able to carry out standard surgeries on anybody better than a team of the best surgeons can do today.
Training algorithms to provide reliable results is a hard and time-consuming job where a human expert still has the final say. The widespread availability of this technology and the ethical scope of application are other issues that specialists are working on.
However, it is hard to deny that technologies are changing many aspects of our lives, including healthcare and social interaction. We are yet to see where machine learning technologies will take us.