Original article can be found here (source): Artificial Intelligence on Medium
What is the difference between Artificial Intelligence, Machine Learning and Deep Learning?
Ever caught yourself wondering what artificial intelligence (AI), machine learning(ML), and deep learning (DL) are and how they are different? Well, let me tell you that you are not alone. These three terms are so frequently used together, dare I say, rather interchangeably that one begins to think they are all the same. No, they are simply not. However, they are very closely related.
The following image shows the relationship between artificial intelligence, machine learning, and deep learning.
From this, we can infer that artificial intelligence is a broad idea and machine learning and deep learning are subsets of it.
Now that we’re sure the three are different, let’s understand them.
Artificial Intelligence (AI)
Coined by John McCarthy in 1956, artificial intelligence is the ability of a machine to think on its own. One can interpret that artificial intelligence, also called machine intelligence, is nothing but computers being able to do tasks that only humans are capable of doing.
Types of Artificial Intelligence
Artificial intelligence can be categorized based on two different factors: it’s performing ability and it’s functionalities.
Types of Artificial Intelligence based on it’s performing ability:
- Narrow AI or Weak AI: Artificial Narrow Intelligence (ANI) or Weak AI is the AI that exists in the world as of now. ANI, as its name suggests, focuses on a single/narrow task. ANI machines are not very versatile or intelligent but they can outsmart humans in their particular task and pre-defined range. The best example is how the bot in a chess game app, when set to maximum difficulty, can easily beat humans unless the human playing against it is Vishwanathan Anand. Siri and Google Assistant are also good examples of weak AI. Though they seem to be too sophisticated and complex to be ANI, their abilities are far from that of humans — they lack consciousness and intelligence to be at par with human beings, hence, they come under this category.
- General AI or Strong AI: Artificial General Intelligence (AGI) or Strong AI is the ability of machines to exhibit human-like intelligence. The AI in movies like Robot (Enthiran) and Her is the best example of AGI. Many artificial narrow intelligence algorithms combine to form an AGI. As of now, though machines are capable of processing data or perform certain tasks much faster than us humans, they cannot do everything that a human being does. This is because we human beings are a lot more than just our ability to comprehend information and take decisions based on our past experiences. Humans possess consciousness and emotions, which are difficult to be tapped into machines.
- Super AI: When people say they are worried about AI, they are referring to Artificial Super Intelligence (ASI). According to Oxford philosopher, ASI is
“any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest”
Machine Learning (ML)
Machine learning, a subset of artificial intelligence, is the study of computer algorithms that learn based on past experiences and improve themselves. Arthur Samuel first came up with the term “Machine Learning” in 1952. Machine learning algorithms learn patterns from the past data, called training data, and build mathematical models. These models can be used to make predictions or take decisions without being explicitly told to do so. Since these predictions are made by the mathematical model based on historical trends, they are likely to be correct.
Types of Machine Learning
- Supervised Learning: In supervised learning, the training data fed to the model includes both inputs and outputs, also known as labels, which helps the model learn patterns from these input-output examples. An example use-case of supervised learning is predicting the sales of a particular product for the next year.
- Unsupervised Learning: In unsupervised learning, no labels are given to the model. The model is left on its own to figure out patterns. As the name suggests, one does not require to supervise the model here, unlike supervised learning where the model is explicitly told the pattern it has to find. An example use-case of unsupervised learning is movie recommendation on Netflix.
- Reinforcement Learning: It is one of the three machine learning paradigms. In reinforcement learning, the reinforcement agent is left in an interactive environment and adapts to the “trial and error” technique and learns from its own experiences. An application of reinforcement learning is to solve different games and achieve great performance, sometimes even better than most intelligent humans.
Deep Learning (DL)
Deep learning is the subset of machine learning, and of course artificial intelligence, that deals with algorithms that work very similarly to a human brain. Deep learning works on artificial neurons, that perform just like the neuron cells in our brain do. Way back in 1943, the first computer model based on neural networks of the human brain was created by Walter Pitts and Warren McCulloch. So, it’s safe to say deep learning finds its roots back in 1943.
I hope I helped you understand the difference between the three in this blog. I will get into details about machine learning and deep learning in the blogs to follow.
Do leave a comment in case of any mistakes in the above content. You can also reach me on LinkedIn or email me at firstname.lastname@example.org.