A simple introduction to AI, ML and Deep Learning

Hello curious minds 👋

Today we will discuss about Artificial Intelligence, Machine Learning and Deep Learning in simplest way I could think of.

Agenda #

Necessity #

The terms AI, ML & DL do surface quite often on the internet nowadays, they are today’s buzzwords. Often people think they mean the same thing but they do not. In this article let us understand what each term means and how do they differ. Now let’s start our discussion with AI.

Artificial Intelligence #

The term Artificial Intelligence (AI) simply means man-made Intelligence. We define intelligence as “The ability to acquire and apply knowledge and skills” and the definition of learning is “The acquisition of knowledge or skills through study, experience, or being taught.” In other words (in terms of Learning) intelligence can be defined as the ability to learn and apply.

Artificial Intelligence is generally categorized into 2 sub-categories, Artificial General Intelligence (AGI), and Artificial Narrow Intelligence (ANI). An AI if it is capable of performing all cognitive tasks humans do, then we call it AGI Similarly if it performs the specific cognitive task at the human level or better, then it is ANI.

Unlike ANI, right now achieving AGI is beyond the human reach although possible in the future. In fact, today such algorithms exist which outperform humans in specific tasks. Some examples of ANI are AlphaGO, AlphaStar, IBM Deep Blue, IBM Watson, OpenAI 5, etc.

Right now whenever we use the term AI usually it means ANI.

Machine Learning #

In order to achieve AI, we need a program/algorithm with the ability to learn with experience and apply this learned behavior to solve specific tasks. These kinds of algorithms are known as learning algorithms. AMAZING, right?

We can define Machine Learning as a collection and study of learning algorithms. Upon given a task T, the learning algorithm improves through experience E. We measure the improvement in performance using metrics P. Some common metrics are accuracy, precision, recall, etc.

Some common tasks T are predicting house prices, classifying images, translating between languages, etc.

For gaining experience E we need two things, one is data another is the training process. Depending on the data we further divide machine learning into 3 parts,

  1. Supervised Learning
  2. Unsupervised Learning
  3. Reinforcement Learning

The training process might differ according to the algorithm based on the task we are trying to complete.

Some examples of machine learning algorithms are Decision Trees, Random Forest, Linear regression, Logistic Regression, Neural Networks, etc.

We select ml algorithm(s) based on the task. But if the task involves any of the cognitive abilities such as speech recognition, object recognition, etc then these ml algorithms do not perform well. So to overcome this shortcoming, Deep Learning is used. Let’s get our fundamentals straight about Deep Learning next.

Deep Learning #

To understand Deep Learning, we will take a short detour to learn Artificial Neural Network a.k.a. ANN.

The notion of ANN is inspired by our understanding of the human brain. A human brain has around 100 Billion computational units and each such unit is called a neuron hence the analogy of using artificial neurons in ANN. (we will discuss in detail about ANN in my next blog.)

In ANN these neurons are arranged in a layered manner, the layer which receives input is known as the input layer similarly the layer which produces output is known as the output layer. The layer(s) which are in between the input and output layer is known as a hidden layer(s). If your ANN has more than 1 hidden layer then that is called Deep Neural Network (DNN). Now back to deep learning,

Simply speaking Deep Learning is study of Deep Neural Networks.

Using DNNs we can solve any task with approximate accuracy, provided enough neurons are present This theorem is known as Universal Approximation Theorem(UAT).

As DNNs follow UAT, their different architectures used to solve almost every core AI task such as Natural Language Translation, Image Object Detection, Speech Recognition, etc. That is the reason they are the center of the AI boom today.

Take away #

If you still didn’t fully understood what we discussed.

The least you should take away from this blog is following 😉,

  • Artificial Intelligence (AI) simply means man-made Intelligence.
  • Machine Learning is the study of Algorithms that achieves AI.
  • Deep Learning is the study Deep Neural Networks, a learning algorithm that is capable of achieving all cognitive tasks.

I hope you are now clear with what is AI, ML, and DL conceptually.

We will talk about Deep Learning in the next blog.

for more such AI related content, click here.

Till then Stay Safe and Stay Curious 🤔😉.

Spread the knowledge
Default image
Swapnil Kumbhar
Deep Learning practitioner | Curious | Anime Fan
Articles: 2