History of Artificial Intelligence (AI) and How it has evolved
There will be no person who hasn't heard of the word AI. We are in the 4th industrial revolution and AI is the fuel for it. Through AI gap between man and machine will be reduced. Any technology has power and control.
Here with AI, humans lose control(machines supersede humans). That means machines gain autonomy(UAV, self-driving cars).
The Gestation of AI(1945–1955)
In 1945, after world war 2, people had freedom in mind and started to think in new ways. In 1948 there was a special innovation happened that affected so many fields. That was the transistor. Following that we got digital computers. In the early 1950s, Turing published a paper called “can a machine think?”. In 1943, McCulloch and Pitts implemented a neural model as the very first intelligent program.
Birth of AI(1956)
Marvin Minsky and John McCarthy were two leading people in introducing AI. They completed a neural network-related Ph.D. at Princeton University. After completing their PhDs McCarthy joined Carnegie tech and Minsky joined MIT. The Dartmouth conference in the USA is a special occasion in the history of AI. During the conference, MacCarthy said that intelligent machines need a separate discipline and he named it Artificial intelligence. According to MacCarthy’s definition, AI is the science and engineering of building intelligent machines.
After the conference, MacCarthy joined MIT to work with Minsky. MacCarthy prompted logic-based intelligence which is symbolic AI while Minsky said that intelligence is not always logical and prompted an anti-logic approach which is later known as ML(machine learning).
Early development of AI(1952–1969)
During this period, logic-based intelligence programs were implemented under symbolic AI such as problem solvers, game players, and theorem provers. In the early 1960s, Minsky criticized the neural network and because of that, there was no development related to the neural network until 1986. At that time, AI researches were mainly funded by the US military. They tried to develop intelligence weapons. There were several reasons for reducing the popularity of AI during that time:
- Since AI research was funded by the military, all the details were kept as secrets. General people didn't know.
- Power and technology both are in hands of AI machines. So, people were afraid of that.
- At that time AI was not a science
Knowledge-based systems(1969–1974)
People noticed that the main ingredient of AI is knowledge. So, they started to model the knowledge for developing intelligent machines. Knowledge modeling comes under symbolic AI which involves logic-related intelligence. Herbet Simon and Allen Newell made a huge contribution to knowledge modeling. Expert systems, Natural language processing, game players, and problem solvers are knowledge-intensive systems.
AI become an industry(1980-present)
It has taken more than 25 years to gain recognition from the industry. Xcon was the very first industrial application of AI. Xcon is an expert system that configures mini-computers to meet user requirements. That process was normally taken 8 hours but xcon did it within 8 minutes.
The neural network was reborn with a backpropagation training algorithm. In the late 1980s AI-based weapons were demonstrated. The best example of that is the DART expert system which is used in the gulf war.
AI becomes a science(1995-)
Until the 1980s AI was considered as engineering-based but there was not much science(unable to explain). After the 1980s AI also adopted scientific methods that are experimental explanations for theories. Being a science AI won a big recognition.
Latest developments
- Alpha Go
- IBM Watson
- Sophia
- Self-driving vehicles
A large component of machine learning is added to the latest developments of AI. Those applications involve training in a large amount of data which are non-algorithmic, and non-symbolic. But early developments of AI were symbolic bases. Future AI applications would be a combination of ML and cognitive systems.
Major universities that are doing AI research
- Carnegie Mellon University (Pittsburgh, PA)
- Standford University (Standford, CA)
- Massachusetts Institute of Technology — MIT (Cambridge, MA)
- University of California, Berkeley (Berkley, CA)
- Harvard University (Cambridge, MA)
- Yale University(New Haven, CT)
- Cornell University(Ithaca, NY)
AI has been influenced by so many other fields
- Philosophy
- Mathematics
- Neuroscience
- economics
- Psychology
- Computer Science
- Computer Engineering
- Control theory and Cybernetics
- Linguistics
Broad Areas of AI
The nature of AI can be easily understood by looking at 2 forms of intelligence in our brains. One form of intelligence is coming from analytical, logical, and symbolic base intelligence. Intelligence related to this area is called symbolic AI or artificial cognitive systems.
And other form is training base intelligence and this comes under machine learning in AI. This is also known as non-symbolic AI.
- Symbolic AI(Artificial cognitive systems)
- Expert systems
- Natural Language Processing — NLP
- Multi-Agent Systems
- Game Playing
- Fuzzy logic
- Ontology
2. Non-symbolic AI(Machine learning)
- Artificial Neural Network — ANN
- Genetic Algorithms
- Robotics
- Computer Vision
This is the end of our article. We have discussed the history of AI and its major areas. Hope this will be helpful for you.
See you in the next blog post. Bye Bye🍻🍸❤️❤️
