Nearly everyone has heard of the limitless possibilities presented by artificial intelligence (AI). Whether it is recognizing your face on Facebook, or communicating to you through Siri, we interact with AI systems nearly every day. So what allows computers to do such complex tasks. The answer is Neural Networks.
To state it simply Neural Networks are incredibly powerful mathematical models that are able to predict data with incredibly high levels of accuracy. They work by taking inputs, randomly producing outputs, and then back-propagating to “Fix themselves.” But how limitless are the possibilities presented by Neural Network? Neural Networks are part of a section of Machine Learning models that use Supervised Learning to teach themselves. Supervised Learning models are only able to synthesize data that they are given and are unable to dynamically adapt to new situations. Hence, even though modern artificial intelligence carries a lot of promise in disciplines such as predicting the winner of the next Superbowl, we are still far from the age where we can communicate with computers as if they are intelligent beings.