Machine learning -- also sometimes called artificial intelligence -- has been a staple of science fiction and futurologists for years. Now, machine learning is finally making it out of the research lab and into the real world – in a way that's easy to implement and cost-effective for all kinds of businesses. Google has been at the forefront of this development and over the course of three posts, we'll explain what machine learning is, how it works, and how organisations like yours are using it.
With conventional IT, you have to provide precise rules to allow a system to complete a particular task: a defective screw is bent or too short or the threads are crooked. With machine learning, the system can work out what the rules are through training, improve its performance over time – and deal with a much greater level of uncertainty and variation in the data, such as the screws being viewed from different angles. Or, in the case of an artificial brain developed by Google, learning to pick out cat videos on YouTube even though it hadn't been fed any information about the specific features used to identify cats. Like the human brain, machine learning can also make predictions – how likely is this to be a cat? – and adapt as it's fed more data, so it gets better at accurately identifying cats.
This kind of machine learning is increasingly being used in business applications. For example, it can automatically figure out how to spot defects during manufacturing – separating good screws from bad ones, for example – or screen medical images to flag up patients who may be showing early signs of a disease. It can understand customers' behaviour in both retail and entertainment in order to predict future demand or suggest other products they might like. It's at the heart of both YouTube's and Netflix's recommendation systems. It's also taking activities such as speech recognition, content analysis of text, images and video, and language translation to the next level.