Quantum Machine Learning—Beyond The Hype

There are two terms that tend to be misunderstood in computer science: machine learning and quantum computing. Let’s discuss these terms in some detail.

What is quantum machine learning?

Quantum machine learning is the use of quantum computing for the computation of machine learning algorithms.

There are many portrayals on these two technologies. They start with machines that understand the natural language of humans, and end with a utopia or dystopia, depending on the media. Don’t fall for the hype! An unbiased and detailed look at a technology helps us steer clear of any misunderstandings that surround it. Let’s start with machine learning.

What is machine learning?

According to Cassie Kozyrkov, Chief Decision Scientist at Google, “Machine learning is a thing-enabler, essentially.”

With machine learning, the aim is to put a label onto an unlabeled thing. There are three main ways of doing this: classification, regression, and segmentation.

Classification

In classification, we try to predict the discrete label of an instance. Given the input and a set of possible labels, we try to guess which one it is. Here’s a picture. Is it a cat or a dog?

Regression

Regression is about finding a function to predict the relationship between input and dependent continuous output values.

For example, given that we know our friends’ income and the effective tax rates, can we estimate the tax rate given our income, even though we don’t know the actual calculation?

Segmentation

Segmentation is the process of partitioning the population into groups with similar characteristics, which are likely to exhibit similar behavior. Given that we produce an expensive product, such as yachts, and a population of potential customers, who do we want to try to sell to?

---