Skip to main content

From Turing to Today: The Story of Artificial Intelligence

 Artificial Intelligence is not the same as it looks today, it took many years to bring the AI to this shape as it sees today, and many researchers, scientists, and mathematicians play vital roles in developing AI to the position where it is now.

Mathematics and AI

AI utilizes many previous scientific inventions not identified here because AI as a science has only existed since the middle of the 20th century. The information below provides examples of key milestones and major trends in AI.

Learn more about the basics of AI in our comprehensive guide: Introduction to AI.

1931:

In 1931, an Austrian mathematician, Kurt Gödel, demonstrated that in first-order predicate logic, every true statement can be derived. When it comes to higher-order logic, on the other hand, there are true statements that can not be derived.

1937:

In 1937, Alan Turing, an English Mathematician and Computer scientist, pointed out the limits of intelligent machines with the halting problem.

1943:

In 1943, McCulloch and Pitts wrote a paper in scientific history, titled ‘A Logical Calculus of Ideas Immanent in Nervous Activity,’ that models neural networks and makes the connection to propositional logic.

1950:

In 1950, Alan Turing again defined machine intelligence with the Turing test and wrote about learning machines and genetic algorithms.

1951:

In 1951, Marvin Minsky, an American Computer Scientist, developed a neural network machine and demonstrated it with 3000 vacuum tubes, which simulated 40 neurons.

1955:

In 1955, Arthur Samuel, an American Computer Scientist, developed a computer program for a game checker that plays better than its developer.

1956:

In 1956, McCarthy organized a conference at Dartmouth College, at which the name Artificial Intelligence was first introduced.

In the same year, Newell and Simon of Carnegie Mellon University presented the Logic Theorist, which was the first symbol-processing computer program.

1958:

In 1958, McCarthy first invented the high-level language LISP at MIT (Massachusetts Institute of Technology). He writes a program that is capable of modifying itself.

1959:

In 1959, an American Computer Scientist, David Gelernter, built the Geometry Theorem Prover, which automatically proves geometric theorems by using established geometric axioms, definitions, and previously proven theorems.

1961:

The General Problem Solver (GPS) was developed by Allen Newell and Herbert A. Simon,n in 1961, was designed to simulate human problem-solving processes and thought, making it a foundational program in the field of Artificial Intelligence.

1963:

In 1963, Professor John McCarthy founded the AI Lab at Stanford University.

1969:

Minsky and Papert show in their book Perceptrons that the perceptron, a very simple neural network, can only represent linear functions.

1972:

In 1972, Alain Colmerauer, a French scientist, invented the logic programming language PROLOG.

Also in 1972, British physician de Dombal developed an expert system for the diagnosis of severe abdominal pain.

1976:

An expert system, MYCIN, was developed by Shortliffe and Buchanan. MYSIN was used for the diagnosis of infectious diseases.

1981:

Japan’s Ministry of International Trade and Administration (MITI), in cooperation with eight leading computer companies, launched a research project, project called the ‘Fifth Generation Project’. The goal of the project was to build a powerful PROLOG machine.

1982:

In 1982, the expert system R1 was developed. It was used for configuring computers, saving Digital Equipment Corporation 400 million dollars per year.

1986:

1986 was a pivotal year; the field of AI experienced a significant resurgence, which was referred to as the ‘Renaissance of neural networks’. It was an expert system that brought advancement in deep learning and the availability of powerful hardware like GPUs, enabling them to tackle complex problems in various fields like computer vision, natural language processing, and drug discovery.

1990:

Pearl and Cheeseman played important roles in the period 1990, they integrated artificial intelligence with probability theory with Bayesian networks. Multi-agent systems became popular in this period.

1992:

TD-Gammon showcased the power of temporal difference (TD) learning by demonstrating the advantages of reinforcement learning.

1993:

In 1993, autonomous robots called Robocup were developed. It was an initiative to build soccer-playing robots.

1995:

In 1995, an American statistician, Vladimir Vapnik, developed a support vector machine from statistical learning theory. That machine is very important today.

1997:

In 1997, IBM developed a chess computer called Deep Blue. Which then defeated the chess world champion, Gary Kasparov.

In 1997, the first international RoboCup competition took place in Japan.

2003:

The robots that performed in RoboCup demonstrate impressively what AI and robotics are capable of.

2006:

Service robotics has become a major artificial intelligence research area.

2009:

In 2009, Google developed the first self-driving car that drove on the California freeway.

2010:

In 2010, Autonomous robots began to improve their behaviors through self-learning.

2011:

In 2011, IBM developed a machine called ‘Watson’ that beat two human champions on the television game show called ‘Jeopardy!’. Watson understands natural language and can answer difficult questions very quickly.

2015:

2015 was the most important year in the evolution of artificial intelligence:

  • Daimler developed the first autonomous truck on the Autobahn.
  • Google self-driving cars have driven over one million miles and operate within cities.
  • Deep learning enables very good image classification.
  • Paintings in the style of the old masters can be automatically generated using deep learning. AI becomes creative.

2016:

In 2016, the Go program AlphaGo, which was developed by Google DDeepMinddefeatedt the European champion with 5:0 in January and Korean Lee Sedol, one of the world’s best Go players, with 4:1 in March. Deep learning techniques applied to pattern recognition, as well as reinforcement learning and Monte Carlo tree search, led to this success.

The AI revolution:

Around 2010, after about twenty-five years of research on artificial neural networks, the researchers began to reap the rewards of this research effort. The deep learning networks are quite capable of learning how to classify images, with very high accuracy in some cases. Since image classification is pivotal for all classes of smart robots, this was the kick-off point for the AI revolution that led to the development of smart self-driving cars and service robots.

Neural Networks

The future of AI is technological advancements, with more money flowing into and changing social perception of AI algorithms and related hardware, which will shape the future of AI. The area ripe for empowerment and development in AI is machine learning.

Comments

Popular posts from this blog

How First-Order Predicate Logic Powers AI Reasoning

Artificial Intelligence, Mathematics, and computer science depend heavily on logic for decision-making, knowledge representation, and automated reasoning. Perhaps the most powerful logical system in all three fields is First-Order Predicate Logic (FOPL). While propositional logic is not able to make very precise statements about objects and relationships, FOPL can. In this blog, we will explore the foundations, structure, and application of First-Order Predicate Logic, along with its syntax, semantics, and real-world relevance. Learn more about the foundation of logical reasoning in AI by reading our post on Propositional Logic. What is First-Order Predicate Logic? First-Order Predicate Logic (FOPL), or First-Order Logic (FOL), is a symbolic formal system that extends propositional logic by introducing quantifiers, predicates, functions, and variables. FOPL enables us to make statements like: ‘Each student has a laptop.’ This is more pr...

Propositional Logic Explained Simply: Learn with Easy Examples

Propositional Logic Artificial Intelligence (AI) uses formal logic systems to mimic human reasoning. Of these systems, Propositional Logic is one of the pillars of knowledge representation and reasoning. Although it’s a basic and well-defined type of logic, it provides an entrance point for grasping more sophisticated logical frameworks in AI, such as First-Order Logic, Description Logic, and so forth. This blog post discusses propositional logic’s syntax, semantics, proof systems, resolution, Horn clauses, computability, and complexity, and its applications are limited in AI. What is Propositional Logic? Propositional logic, also referred to as propositional calculus or sentential logic, is concerned with propositions, i.e., declarative sentences that are true or false but not both. It does not include variables and quantifiers, unlike predicate logic. Propositional logic, in the case of AI, is applied to represent basic knowledge and deduce new facts based on current facts with the a...

The Role of Knowledge-Based Systems in Artificial Intelligence

 Artificial Intelligence (AI) has made tremendous progress over the past decades, from rule-based systems to powerful machine learning systems. Along the way, Knowledge-Based Systems have led the way in embedding human knowledge into computer systems. Knowledge-based Systems try to mimic the decision-making powers of human experts, providing solutions across different fields, from healthcare to finance. Understanding Knowledge-Based Systems A knowledge-based system is a computer application that uses a knowledge-based regarding a particular subject to address complex problems. Unlike conventional programs that use a sequence of instructions, knowledge-based systems use a knowledge base and an inference engine to mimic human thought processes. Knowledge Base : It contains the domain-specific facts, rules, and heuristics. Inference Engine: Uses logical rules on the knowledge base to derive new information or make decisions. The structure enables a knowledge-bas...