By William O., Age 15

     Often online shoppers will encounter “recommended for you” when buying items online at sites like Amazon, music services, and even news articles.  How do they decide which items to target and which ads to use? Word suggestion in texting and typing is also common form technology that people encounter every day. How does the program know what word you are trying to type? The answer is machine learning.
     Many people confuse artificial intelligence with machine learning, but they are not the same thing.  Artificial intelligence is merely programming a machine. Most conventional programs use artificial intelligence.  However, machine learning goes a step further allows the code to be able to respond to new input and stimuli to create new and different responses.  Machine learning is more common in professional uses than mere artificial intelligence because it’s higher applicability. For example, robots are an example of artificial intelligence but not machine learning because they don’t really respond to new input, they just run their existing code.  On the other hand, Apple’s “Siri” and Amazon’s “Alexa” use machine learning in both speech recognition and language processing to understand the input of what owners are asking of them. 
     The basic summary machine learning is programming a machine to learn.  There are several ways of doing this. One of the most basic and often used for teaching purposes is the genetic algorithm, which is creating several miniature programs, putting them through tests, then making more programs based on which miniature one did the best on the tests and destroying the mini programs that didn’t perform as well.  This is designed to mimic genetics in real life.  
     Another way of programming machine learning is neural networks.  These try to mimic the human brain inside a computer by replicating individual neurons and their connections.  This is used for predicting and extrapolating from databases. For example, it will group a lot of products by seeing which items are frequently bought together and placing those items items in the same group.  If a consumer buys one item in the group, it can recommend other products in the group.  
     

     Artificial intelligence and machine learning are important for the future because they are much faster ways to allow programmers to make high-level prediction algorithms for a variety of areas.  For example, doctors can look at symptoms of new patients and match them with the symptoms of patients that have already been diagnosed to match with the type of disease that they have for faster diagnoses in health care.  Machine learning will also be used in self-driving cars to recognize traffic signs and signals and adapt to increased traffic patterns during rush hours.