The world has seen four industrial revolutions - the first in 18th century triggered by mechanization, such as steam engine, the second in 19th century based on electricity, the third in 20th century by computers, digitalization and the Internet. At the center of the forth industrial revolution we are going through, in 21st century, the driving forces are Artificial Intelligence, Machine Learning, Data science and other data centric technologies. We can say that data is now the new oil.
There are many associated terms such as data science, machine learning, and deep learning. Here I will give a brief introduction to the field, and you can check my Medium/GitHub/LinkedIn pages for a more detailed and technical account.
Before we talk about Artificial Intelligence or AI, it is important to understand what intelligence is. This is a very hard question, as hard as what is life? You may think it is easy to define life since we have so many types of life around, but that is not the case. Every other day, researchers come across examples that defy any definition of life. So far, the acceptable definition we have is 'Life as we know it on earth.' So we create a definition from the example and not the other way around. Due to the difficulty in defining what life is, it is very hard to know if there is any other life - extraterrestrial life - in the universe. Note that it is entirely possible that there is life out there in some other place, but it is not like life on earth, not based on Carbon/DNA/RNA. For intelligence, we mostly have one example - that of human intelligence - so we will define intelligence with respect to human intelligence. Note that again, there may be intelligence but not like that of humans, and we do not know about it.
Artificial Intelligence works on the premise that it is possible to build human-like intelligence out of machines! In fact, this is what was proposed by Alan Turing around the 1950s. There are many ways human beings can exhibit their intelligence, for example, by building a paper clip or a jet plane, but the simplest and most common way is to use language. Now we are more or less sure that any kind of intelligence must have abilities related to language (speech, read, write, process, understand) at the center. In fact, one of the greatest achievements of human beings can be considered the global network they have created, made possible by language and communication. Human beings can pass on very complex ideas to other fellow humans using language. Turing proposed that if a machine can answer questions like human beings, then it can be considered intelligent, and this type of intelligence is called Artificial Intelligence. ChatGPT is an example of that.
A truly intelligent system must have many more capabilities than just that of consuming, understanding, and creating language. Pioneers of Artificial Intelligence, Stuart & Russell, mention the following:
- Language - Natural Language Processing
- Knowledge Representation
- Learning from Experience - Machine Learning
- Sensing and Manipulating the Environment - Robotics
There may be a few more, but the above are sufficient for our discussion. Let me come to the third point, which is about learning from experience or machine learning.
Machine Learning is confined to creating computer software that will learn from data without being explicitly programmed and improve their performance in certain tasks, such as understanding speech, recognizing images, or making predictions. This ability to learn from experience (data) is one of the key defining features of human intelligence. Currently, we have many systems such as Apple's Siri, Amazon Alexa, or ChatGPT which are based on machine learning.
Note that there are many different algorithms used in machine learning for different tasks, and you can read more about them on my Medium page.