What is it in the deep learning that a tech giant buys startups for so much money? The deep learning is one of the fastest growing branches of artificial intelligence. It involves creating neural networks, i.e. such information systems, whose design and function resembles the work of the human brain.
The neural network is an enormous number of interconnected and cooperating processors running at the same time. Each of them has an access to local storage and is supplied with large amounts of data and information on the relationships between data. For a neural network to “learn” well, the software must show how to behave in response to specific external stimuli (i.e., entering data by a computer user).
The deep learning, or rather the idea of neural networks, has existed for fifty years, but it was not very popular until late nineties because of insufficient data available. What made the situation change in less than two decades? First, we have a lot more data needed to build a multi-layer neural network, and second, with the latest technology, we now have machines with much higher computing power.
The human brain contains an average of about one hundred billion neurons, and each one is connected to another ten thousand neurons, so the number of connections between neurons – synapses – is… from 100 trillion to 1000 trillion. That’s a lot. So far we are not able to build a neural network of this size and with such high computing power (though Google is working on the creation of artificial neural networks comparable to the brains of laboratory mice).
To mimic the human brain as best as possible, researchers create complex neural networks spread across multiple layers, and that multilevelness leads to deeper learning. It is worth mentioning that the deep learning is more like teaching a little child than traditional programming, because it is based on cognitive processing. This allows machines to „understand” human signals and „respond” in a human-readable way.
Deep learning is based on the working mechanism of the human brain – network connections form tangled layers on many planes. Teaching the system involves „updating” the connections after each new stimulus. A machine based on deep learning imitates the behavior attributed to people – imitating voice, identifying images, or forecasting.
Deep learning vs. machine learning
There are a vast number of articles on artificial intelligence and the fact that we should seriously consider its potential. Every one of us has probably also found information on both machine learning and deep learning. The nuances between these three technologies are obvious to experts. What about the rest of us? What are the biggest differences between machine learning and deep learning, and can one say that one has an advantage over the other?
Let’s start with the weaknesses of machine learning:
- it requires the presence of a human who, by introducing thousands of examples, teaches the machine how to draw knowledge from them
- all errors must be corrected manually
- the learning process itself is very time consuming
- the machine is fully human-dependent, so it is hard to determine its level of intelligence
Deep learning happens rather without human supervision and control. I said „rather” because there are known cases where deep learning requires human presence. However, for the most part, deep learning is the creation of neural networks that allow the computer to think independently (without human participation).
The power of the deep learning model comes from data. It needs over ten thousand data cases to achieve a good result. Of course, this number is not the only requirement, since the data must involve many variations for the model to understand that the change is not correlated with the class.
Artificial Intelligence is like an umbrella that covers both deep learning and machine learning. If we were to draw a circular diagram illustrating the relationship between Artificial Intelligence, Machine Learning and Deep Learning, then the largest circle would be reserved for artificial intelligence, with a “plot” for machine learning in the middle, and at its very center – for deep learning.
Deep learning in e-commerce
Do you think that the deep learning is still a technology that is not commonly known? Quite the opposite. Google’s translator is using deep learning technology. Moreover – two years ago, Microsoft had created the deep learning algorithm that is more accurate than humans. During the test, it classified faultlessly over three hundred thousand images into one of the three hundred descriptive categories.
And how can e-commerce benefit from the deep learning? First of all, it will be much easier to get even more accurate demographic data, because computers are constantly „taught” to differentiate age and other indicators. In addition, traffic analysis in the store can be more detailed, since you now have more knowledge on the activities of a potential buyer in your store; who they are, and what they do on the site.