ChatGPT, Should We Embrace It?

ChatGPT Introduced Itself as a More Advanced Type of AI Compared to Others.




Hi, everyone.

How are you guys?

For me,

upcoming group teamwork and presentation events have led me to interest in Chat GPT.

It is hard to deny that it is a controversial software. 

And even some experts are saying that it will dynamically affect the IT industry. 

Since it is based on AI technology, 

I will introduce some interesting history and facts about it.




The History of AI


Did you know that AI was first proposed in the 1950s? 



                                                 Image from Wikipedia

                                                                                               Alan Turing (1912 -1954) 

                    a mathematician, also known as the main character in the film called The Imitation Game (2014)



Artificial Intelligence




British mathematician Alan Turing described how to test whether a machine can think, the
 
possibility of developing an intelligent machine, and the machine to learn in a paper called

"Computing Machine and Intelligence".





Perceptron




Meanwhile, Warren McCullonch, an American neurosurgeon, and Walter Pitts, a logician,

have theoretically demonstrated that connecting artificial nerves with basic functions that

turn on and off like an electric switch can mimic very simple functions in the human brain.


HEP (High Energy Physics) also proved that when repeated signals occur within biological 

neural networks, neurons have a kind of learning effect to remember the signals.


These studies had a decisive impact on the work of Cornell University psychologist Frank 

Rosenblatt in 1958, which led to the birth of Perceptron (an artificial neuron that simulates 

brain nerves). 



                                                                                                              Image from Data Analytics

                                                                         Figure of how Perceptron works



As a result, a revival period of neural network-based artificial intelligence research gained 

popularity.

(You can see that the concept of deep learning was already proposed 70 years ago.)




Machine Learning 


Since the 1970s, most companies have focused their R&D direction on practical statistical 

technology.

They have shown that experimental planning and statistical analysis techniques can affect the 

improvement of manufacturing quality/production efficiency.

Under the name of data mining, this concept has been used to address inefficiencies in the 

industry and has become the foundation of big data technology.


                                                                                                                  Image from WallStreetMojo




Artificial intelligence research spread in the 1980s with the introduction of an expert system 

called "Machine Learning" in the industry. 

It uses algorithms and neural network models to assist computer systems in progressively

improving their performance. 

Machine learning algorithms automatically build a mathematical model using sample data – 

also known as “training data” – to make decisions without being specifically programmed to 

make those decisions.


Therefore, groundbreaking AI studies such as 

back-propagation, a basic model of deep learning, has been published.

However, have disappeared without being noticed by the world 

due to various limitations such as computer performance, limited utilization, and machine 

learning algorithms.





Deep Learning


When it comes to the 2010s, the algorithmic change was not that different from the previous 

models. 

but the difference was the amount of data capable.

due to the Deep Learning, 

machines began to recognize the world, 

identifying and distinguishing the information.

Think about Face ID in iPhone  and self-driving car. 

                                                                                                         Image from Semiconductor Engineering



Chat GPT

and then Chat GPT appeared last November 2022,


The term Chat GPT stands for "Chat Generative Pre-training Transformer".

Meaning that it pre-learns all the texts in the digital world 

and makes a probabilistic map of language in every human-made sentence. 

When I input some words into Chat GPT, it generates output with new sentences.

If AI has been identified and distinguished over the past decade, 

it is Chat GPT that  showed artificial intelligence can generate nowadays.




Conclusion 

The idea of deep learning, the process of how AI works were already been proposed since 

1950s, now due to the emergence and development of new technology enabled to process 

the enormous amount of data.

It led to the realization of computers identifying existing information.

Now, it is Chat GPT that embraced the capability of computers to generate and 

visualize such information.




The post is getting to long so it will be continued on next posts

dealing with contents of focusing more about the software Chat GPT.





Reference

Foote, K. D. (2023, March 28). A brief history of machine learning. DATAVERSITY. Retrieved April 6, 2023, from https://www.dataversity.net/a-brief-history-of-machine-learning/ 




Comments

Post a Comment