Skystis Infotech LLP

Skystis Infotech LLP / Machine Learning  / Machine Learning (ML) 

Machine Learning (ML) 

When you hear the word Machine, what is the first image that pops into your brain? A computer? A car? An ATM? From the Cell Phone in our hands to the electric lines running above our head, everything is a machine. In today’s world, almost all of our work is dependent on machines, Machines have become a very essential part of our day to day life and yet we are unknown to its utmost potential. Every day there are new discoveries made in the field of Technology, from high-speed internet to data-based voice recognition, from maglev trains to self-driven cars and so much more. Until recent times we have programmed machines to do stuff, we have prepared scripts for it to read and perform accordingly. But what if these machines starts learning this stuff itself? What if we no longer need to write those long programs for it to perform? Well it’s no longer a vision, it is a reality and we call it Machine Learning

machine-learning.jpg

Machine Learning is a method of data analysis that improves automatically through experience and by the use of data. It is based on the idea that systems can learn from data, identify the pattern and make decisions without being explicitly programmed to do so. Machine Learning is a study of computer algorithms that improves automatically. It is a branch of Artificial Intelligence

Some of the places where you might have seen the application of machine learning are as follows:

  • Wearable Healthcare Devices which monitors your health through sensors, sometimes this technology can even help you prevent the unfortunate
  • The list of recommended items you see is based on your previous searches or the pages you have visited, these websites are using machine learning to analyze your shopping history.
  • financial businesses use machine learning to prevent fraud, they can analyze the client’s history to detect risky profiles.
  • government can use machine learning in different sectors as it itself has multiple sources of data.
  • machine learning technology is used in fully automated cars to analyze and detect the upcoming obstacle
  • the voice recognition or facial recognition on your phone is based on machine learning

In 2014 a problem with tesla vehicles was diagnosed where components of the engine were overheating, using this data every vehicle was automatically repaired by a software patch. All the tesla vehicles whether or not they are Autopilot enabled- send data directly to the cloud. This is the data that is used to build the next-level automatic car, with internal as well as external sensors placed in such a way that it can detect the driver’s hand placement on the instrument and how they are operating them. This data is used to improve its system and can also be sold for billions of dollars.

The future of machine learning is quite bright, we can expect an unthinkable revolution in the next 10 years. Even today the error percentage of computers has gone as down as 3%. Experts believe that after a decade or so there will come an era where machines will affect all the professions like doctors, engineers, accountants, drivers, microbiologists, etc. An expert of machine learning Jeff Dean believes that in the future there will be one big model that will be able to do plenty of things, from diagnosing your diabetes to be used in cars and drones all at the same time, with this one big model there will not be a need to create small models that does some specific tasks. The advancement in Artificial Intelligence is such that we can now say that “all that is imaginable is achievable”.

Nothing in this world is perfect, with the pros comes the cons, and such is with machine learning also. We already live in a world where there is a very thin line between reality and science fiction and this line is erasable by the extreme development of artificial intelligence. What was once sci-fi entertainment content is now a possibility. What makes machine learning a growing risk is the ability to make complex decisions itself. They work on making a prediction based on previous data and thus there is a possibility that some of these decisions might not fit the future situation and be considered wrong but the price of that wrong decision can be fatal. What if an automatic car made a wrong decision and caused a horrible accident, or what if a self-performing surgery machine made a wrong decision and caused the ill fate of the patient, or what if a trading software based on machine learning caused you great financial loss.

There is no way one might know what will happen but there are ways in which we can always take precautions. Before entering the machine learning system in the market we should try to run it in a controlled environment where it can function freely without any human intervention but still ultimately be in our control. In these simulations we can see how it reacts under different situation and how it evolves accordingly, with changing the environment around it we can also observe how does the system handle and tackle that change and grow with it. This way we can control the risks to some extent.

Machine learning has great potential in the upcoming market. But with this technology the risk it poses will increase and thus it is very important for the creators to predict and work on the solutions of the problems before they arise so that the balance is maintained and no harm is done.

No Comments

Post a Comment