The Evolution of AI

Have you ever heard about AI? Well, you're in luck as I will tell you exactly that!

Artificial intelligence, or AI is the simulation of human intelligence. A machine can be classified as an AI if it possesses human-like intelligence. Nowadays, Chess Machines have evolved to such an extent that their skill exceeds that of any human, AI chatbots are replacing typical consumer helpers. But, how did we get here?

It all began in the mid-20th century when AI was stuff for fantasies. In the background, this was being made a reality by a group of visionaries, including Alan Turing.

Alan Turing, a British mathematician and logician devised the Turing Test, a simple yet integral idea in the field of AI even today; If a machine could carry on a conversation such that a human could not distinguish it from another human, the machine could be considered "intelligent".

1st Stage of AI: Rule-based AI

The first stage of AI was Rule-based. In this, machines were explicitly

programmed to carry out certain tasks, nothing more, nothing less.

For example, if you wanted the machine to identify a picture of a dog you'd have to code all the characteristics of a dog into it.

if 
 var tail== true{1}
 var ears== true{2}
 var ears== true{2}
 var legs== true{4}
 var fur== true{unquantifiable}
dog== true
else:
 dog== false        

They were impressive for their time but lacked the flexibility and adaptability of human learning.

2nd Stage of AI: Machine Learning

Then came Machine Learning, which revolutionized AI forever. It allowed AI to learn from data and experience rather than explicitly coded programs.

For example now you could feed data into the machine about dogs, like pictures, information, etc, and it would identify the picture of the dog more efficiently and accurately.

<dev> transfer {dog1.jpeg} to database
      transfer {dog2.jpeg} to database
      transfer {dog3.jpeg} to database
      transfer {dog4.jpeg} to database
      transfer {dog5.jpeg} to database
      transfer {dog_info.txt} to database
      transfer {dog_info2.txt} to database        

<USER> [animal.jpeg] Is this a dog?        

<MACHINE> Processing......
          Crosschecking with database......
          AFFIRMATIVE!        

However, the dawn of AI was yet to come...

3rd Stage of AI: Generative AI

However, this wasn't the end as a new subfield of ML(Machine Learning) emerged, Deep Learning, It was inspired by the human brain's Neural Networks.

They contained several layers of neurons which allowed them to process vaster amounts of data faster than ever! They could perform various tasks such as Natural Language Processing, biometric recognition, voice recognition, etc.

Now these machines could generate content! For example, now you could ask the machine to generate a picture of a dog!

<User> Generate a picture of a dog.
<Machine> Generated! [Click to Download]        

Natural Language Processing

Natural Language Processing is the concept of AI generating human-like text.

Let's take an example of Siri, iOS's personal voice assistant:

If you ask Siri a query, it converts your voice into text, consider it like your phones text-to-speech feature but vice-versa; it then processes the query, produces the output and converts the output text-to-speech!

Generative Adversarial Networks

Then came GAN frameworks, comprising 2 neural networks, the generator and the discriminator.

The generator generates the data which the discriminator then reviews.

For example if the user gives a prompt to the AI to generate the image of a dog, the AI generates the picture of a dog and sends it to the discriminator for cross checking. If the generated output doesn't match the predicted output, the AI tries to get the generated output closer to the predicted one; this is a classic example of an optimization problem. It can graphically be represented as shown below:

Although it may seem like this is the end, it isn't. New developments in AI are being made every day, probably at this very moment! No matter we achieve, there's never an end to human exploration.