It may seem to you that machine learning (ML) and deep learning (DL), which includes neural networks, are the same thing. But this is not true. Strictly speaking, DL is a subset of ML.

In the case of "classical" machine learning, selected features of the objects under study are fed to the input of the model to obtain predictions. In deep learning, the neural network receives "raw" data (for example, images) and it itself tries to extract important features in order to solve a specific problem.

What about Apple?

Obviously, the iPhone and other Apple products use neural networks. But why do Tim Cook and other company representatives call them the more general concept of machine learning from presentation to presentation? I don't work for Apple, but from my experience watching people from Cupertino, I can guess the root cause.

How are smartphone cameras similar to human vision?
The cameras of modern smartphones are very similar to our vision and brain. In this article, we will explain this similarity.

You might think that Apple is always playing catch-up. Even if we take the topic of AI, then "their own ChatGPT" was presented by everyone who is not too lazy. Apparently, the company from Cupertino is lazy. But the point is different: Apple always tries to make products that it is confident in, and rarely chases hype. And this reveals the main reason why the company calls neural networks machine learning.

"Classical" machine learning methods are considered more interpretable and valid because they are mostly based on mathematical models. Neural networks are also a kind of mathematical models, but they are very difficult to explain.

For example, banks often use decision tree-type models to solve the credit scoring problem (predicting whether to give a loan or not). This is due to the fact that they are simply interpretable: a decision tree can be drawn, and you can understand how it works. This will not work with neural networks.

Summary

In summary, I'm guessing that interpretability and machine learning validity are major factors for Apple when choosing titles. The company from Cupertino is very careful when it comes to neural networks and everything that may carry some risks. And although they actively use deep learning approaches, they continues to call them machine learning within the framework of marketing.

Share this post