Machine learning is the field of study that “gives computers the ability to learn without being explicitly programmed” (Arthur Samuel, 1959). It’s a type of AI (Artificial Intelligence) in a sense that we let the machine mimic human behavior such as “learning” and “problem solving”.
In 1997, Tom Mitchell gave a concrete definition that has proven to be useful in any ML projects.
“A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.” — Tom Mitchell, Carnegie Mellon University
It can be used to create an algorithm to predict future events based on previous data, for example, stock price prediction (task T), you can run it through a machine learning algorithm with data about past stock prices of a company (experience E) and, if it has successfully “learned”, it will then do better at predicting future stock prices (performance measure P). How cool is that? With the right tools and algorithm equipped with good analytical, you can earn some bucks.
That’s an example for investment professionals and those people involves in stock trading. But Without us knowing, machine learning plays an important role in our daily life like stopping computer malware and email spam filtering. Without machine learning, you will end up frequently going to a computer repair shop for virus removal or you will spend hours searching for your one important email message out of hundreds of spam email. Imagine those numbers of hours we might end up wasting without machine learning.
Because of its capability of solving complicated problems, machine learning evolves rapidly from email spam detection to complicated but useful applications such as improving search results (Google) and self driving cars (Smart Cars). As of December 2016, this type of smart card is still being developed by Google X. That’s awesome and it makes me excited about the future.
Aside from those applications, there are still many machine problems left unsolved. Examples of machine learning problems include, “Is this cancer and how to cure it?”, “Is there any other earth-like planets in the universe and where exactly to find them?” and “Can we create a robot that can be as outsmart us? (e.g. a robot that can study engineering and can program themselves and other robots to improve and educate themselves)”.
It’s very clear that Machine Learning is an incredibly powerful tool that can solve some of our most pressing problems. It can also open up our world to a bunch of opportunities.