Hey Starlord here,
Welcome to the next chapter on AI . I hope my last two chapters give you an understanding of AI and its history. Now it’s time to talk about some technical concept of AI. Let’s start with Deep Learning!
Deep learning is a form of artificial intelligence that uses a type of machine learning called an artificial neural network with multiple hidden layers that learn hierarchical representation of the underlying data in order to make predictions of given new data.
Now, that’s quite a lot of information to absorb, so let’s take one step back and break things down so that we can understand exactly what each of these terms mean. First, let’s see how artificial intelligence, machine learning, and deep learning all are relates to one another.
Image source
As we can see, deep learning is a type of machine learning which is a type of artificial intelligence. Essentially, there are several ways we can create artificial intelligence. Machine learning is one of those ways. In addition, there are several types of machine learning.
Image source
Deep learning is a specific type of machine learning. So, what then is artificial intelligence? As I already told you in my first chapter of Artificial Intelligence is a field of computer science that attempts to create machines that act rationally in response to their environment.
This can be done using variety of methods. First, we can explicitly program a computer to execute tasks step by step. This is how artificial intelligence was first implemented with varying degree of success. However, this technique only works for problems that a programmer can describe as a step by step set of instructions that a machine can follow.
In addition, this technique only works well in constrained environment. It doesn’t handle novel situation very well. Next, we can encode domain knowledge into a machine so that we can use that knowledge to complete tasks.
This technique is used create what are known as expert systems: a database of facts and rules in an interference engine that can deduce new facts given the known facts and rules. These systems have had limited success in various decision support areas, such as medical diagnostics, logistics, and business applications.
Image source
Finally, we can create artificial intelligence by teaching a machine how to solve a problem by identifying statistical patterns in data; a type of artificial intelligence known as machine learning. So now, let’s talk what machine learning is and how it works.
Well, machine learning is a type of artificial intelligence where we teach machine how to solve problems without explicitly programming them to do so. Instead, we teach them by example using data. Machine learning is the application of statistics to the problem of artificial intelligence.
We’re teaching machine to learn how to solve a problem by identifying statistical pattern in data. In essence, with machine learning, we use existing data to learn a function that can make prediction for given new data. For example, imagine we want to create a function to determine whether a photo contains an image of a cat or not.
First, w need to create a dataset that contains images with cats, and images without cats. We’d have a human being label each photo indicating whether t contained a cat or not. Next, we’d apply a machine learning algorithm to the data set of images with and without cats.
The algorithm would learn a function that predicts whether an image contains a cat or not. Finally, if we’ve done everything correctly, we should be able to provide the function a new image, and it should tell us whether it contained a cat or not. Just like we teach a newly born child.
Image source
Now, while this is a vastly oversimplified explanation of machine learning, it captures the essence of what we’re attempting to accomplish. There are numerous machine learning techniques and algorithms in existence today. We have algorithm for classification regression, clustering and many more.
More than I could possibly cover in a single chapter. However, in order to understand deep learning, we need to learn about specific type of machine learning called neural networks.
So, in next chapter we’ll talk about what an artificial neural network is, and how it works. Till then Keep Learning
