Artificial Intelligence or AI is about developing intelligent machines, which can act and behave like a human. Artificial intelligence or AI is the wide-ranging branch of computer science that ranges from general learning and perception to specific, such as playing Chess, driving cars, proving mathematical theorems, diagnosis the disease, etc. It Imparts the ability to think and learn on the machines. Artificial Intelligence (AI) can also be said as anything that makes a machine to act more intelligently. In this tutorial on Artificial Intelligence (AI), we will discuss the definition, types, applications, and history, etc.
AI is the combination of multiple fields of studies hence also known as a multi-disciplinary field. It combines various fields such as:
- Computer science & electrical engineering: It determines how to implement AI in software and hardware.
- Mathematics & Statistics: The mathematics & statistics determine the applicable models and measures the performance of models.
- Psychology & Linguistics: These two fields also play an important role in understanding how AI works.
Note: Artificial Intelligence is not about replacing human experts but extending human capabilities that machines could not do on their own.
Definition of Artificial Intelligence (AI)
Since the time AI is developed, there have been many contradictions between its existence and what it is. As it has evolved, the definition and meaning of AI have changed over the year. There have been several attempts to define AI by different people. Some basic definitions of AI are:
- “It is the ability to learn and solve problems.”—Webster’s Dictionary.
- “AI is the intelligence exhibited by machines or software.” — Wikipedia.
- “The AI is the science and engineering of making intelligent machines”—John McCarthy.
So, in conclusion, we can define it as: “The artificial intelligence is about developing machines that can learn from experiences, solve the given problem efficiently, and performs the human-like tasks.”
What is Artificial Intelligence Exactly?
When we try to understand AI, the questions such as “What is intelligence”, “How does our brain work”, or “How to measure the intelligence”, all these questions are meaningful and also helpful to understand the concept of AI. As humans, the most important and different thing that we have is our mind or intelligence, which provides us the capability to think, perceive, understand, act intelligently, etc. If a machine can have all or some of these intelligence capabilities, that would also be called intelligent. Artificial intelligence is not just to understand intelligence but also to develop intelligent entities or machines.
In the year 1955, the great mathematician, John McCarthy, coined the term Artificial intelligence and defined it as:
In the beginning, people define it as whatever machines cannot do yet, but humans can, is the true artificial intelligence, such as Playing Chess, recognizing images, etc.. Still, as it has developed, the meaning of AI has changed. So, at this point of time, when it is being developed so far, we can generally define it as:
“Artificial Intelligence is the field of study to develop intelligent machines that work and act like humans.”
The AI development currently includes seeing like humans such as Computer Vision, Listen like a human such as Speech recognition, etc.
Till now, AI is defined are based on the thought process and reasoning for the basis of behavior. Norvig and Russell have given four approaches to define the complete field of AI; these approaches are:
- Thinking Humanly
- Thinking rationality
- Acting Humanly
- Acting Rationally
The first two approaches are about the thought process and reasoning, and the other two deal with the act & behavior of machines.
The first approach to define AI is thinking humanly, which defines the cognitive modeling approach. According to this, if it is said that a machine thinks like a human, so first we should understand how humans think through their thought process or through observing the brain in action.
This approach follows the “laws of thought process”. The Greek philosopher Aristotle first tries to code the reasoning process using the Syllogisms laws. These laws of thought somewhere simulated the operation of the mind and generated the new field of reasoning, known as Logic.
Acting humanly means that machines can perform the task the needs intelligence. An example of this approach is the Truing test, proposed by Alan Turing in 1950. According to this test, if the human interrogator cannot tell whether the responses come from a person or from a computer, then the machine will be said intelligently. To pass this test, the machine should possess the following capabilities:
- Natural Language processing
- Knowledge representation
- Automated reasoning
- Machine Learning
- Computer Vision
All these six capabilities compose most of the AI-enabled machines.
The acting rationally approach deals with rational agents. The rational agent is the computer agent that acts in order to achieve the best outcome.
Intelligence in Artificial Intelligence
Before going in-depth to understand AI, we must understand what intelligence is, its components, and how human intelligence is different from Machine intelligence.
What is Intelligence?
Intelligence is the ability to perceive the information, retain its knowledge, and apply the Knowledge to demonstrate the adaptive behavior within an environment or given context. More generally, it is the ability to think, understand, learn from past experiences, apply logic to solve a problem, reason, and to adapt to new situations.
The intelligence is made-up of the following components:
- Reasoning: It is a set of thought processes that enables us to make decisions, predictions, and make a judgment.
- Learning: It is the process of gaining knowledge and information by studying, practicing, past experiences, visualizing, etc.
- Problem-solving: It is the ability to perceive a problem, identifying and selecting the best alternatives for the desired solution.
- Perception: It is the way to acquire, interpret, select, and organize sensory information.
- Linguistic Intelligence: It is the ability to understand the written and spoken language, as well as the ability to speak and write also.
Types of Intelligence
As per the researches of the human brain, there are different types of areas that correspond to certain knowledge and independent of each other. These areas are divided into eight types of intelligence, which are given below:
- Logical-Mathematical Intelligence: The ability to understand and know the answer to certain problems before it is visualized. This intelligence is responsible for the problem-solving ability of someone. Most Mathematicians and Scientists have this intelligence.
- Linguistic Intelligence: The ability to speak, talk, recognize, read, writing poems, learning new languages, and use syntax, semantics, and phonology(speech & sound). Most of the Narrators have this intelligence.
- Spatial Intelligence: The ability to think in three dimensions. With this intelligence, people can easily solve spatial problems such as drawing, reading maps, solving maze problems, construct 3D images, and rotate them. Astronauts, Physicists, and Map readers have this intelligence.
- Musical Intelligence: The ability to understand the meaning made by sounds, understanding pitch, rhythm, sing, and can easily memorize the songs and tunes. Singers and Musicians have this intelligence.
- Bodily-Kinesthetics intelligence: The ability to use the whole body or a part of the body to express ideas & feelings, and the use of hands to transform elements. Players, sportspeople, and Dancers show have this intelligence.
- Interpersonal Intelligence: The ability to know oneself best on its own. By this, one can distinguish their own feelings, set goals, and self-motivated.
- Intrapersonal Intelligence: The ability to understand and recognize other’s feelings, beliefs, and intentions. Mass communicators, Interviewers, etc.
Note: A system or machine can only be said artificially intelligent when it shows or demonstrate at least one or at most all the intelligence discussed above.
Artificial Intelligence Disciplines
AI is a combination of various disciplines, which contribute to building an AI system. The AI system cannot be created by using one technology; rather, it is the combination of multiple fields, hence called an Interdisciplinary field. These disciplines are given below:
- Computer engineering
- Control theory and cybernetics
- Big Data
Philosophy plays an important role in AI, as it deals with answers to various questions, such as “Where does knowledge come from?”, “How does knowledge lead to an action,” etc., it helps to analyze the connection between Knowledge and action, which is very important for the AI system, as the intelligence needs action as well as reasoning.
Mathematics provides us various tools to manipulate the logical certain and uncertain statements and probabilistic statements. It also helps to understand the computation & reasoning about the various algorithms.
The economics field helps to formalize the problem of decision making, and hence it maximizes the expected outcome for the decision-maker.
Neuroscience is used to discover the facts about neurons and how our brain works. It also helps to understand how it is similar to and different from a computer.
Psychology is a field to study and understand the thought process of a human or animal. It helps the AI to understand the human brain, behavior, and acts, which is important to simulate the working of the human brain.
Computer engineering is essential to implement ideas in the real world using codes and programs. It helps to update the values or properties of the AI applications based on the provided data.
Control theory& Cybernetics:
The control theory & cybernetics explains how the artifact(an object made by a human) can operate on their own control.A water clock with a regulator that maintains a constant flow rate was the first self-controlling machine.
Linguistics helps to understand the relationship between language and thoughts. The modern linguistic and AI together form a hybrid field known as computational natural language processing.
Big data is a type of fuel for the growth of AI, as it provides the platform and tools to process and manage huge datasets. To build the machine learning model and deep learning understanding, we require a huge dataset, which cannot be saved or processed by the old data processing tools; for such cases, we need Big data.
How does Artificial Intelligence (AI) work?
Before understanding how AI learns, let’s first understand what intelligence is?
Human beings have inbuilt intelligence, which governs every activity in our body.
The machines do not have any inbuilt intelligence and only have that intelligence that we give to these machines. It means we enable the machines to examine examples and develop the machine learning models based on the given input and desire output.
Machines or AI entities learn in three ways, which are given below:
- Supervised Learning: Machines Learn with Supervision using the labeled datasets.
- Unsupervised Learning: Machines Learn without any supervision using the unlabeled datasets.
- Reinforcement Learning: Machine or agent learns with the actions & feedbacks.
Why Artificial Intelligence (AI)?
Artificial Intelligence has now become the need of human beings rather than technology. If we saw any machine around us, we will find that almost everything is running through AI technology. So Why AI has become so crucial for the human being? There are some reasons for the demand for AI is as follows:
More Data storage:
We are generating more data every day through social media, IoT devices, etc. So, AI is a solution to process more data and derive useful insight, and grow business with the help of data. Big data helps to recommend a better product while shopping online and classify an object from an image. AI is trained on large data sets, and Big data helps us to do this more easily.
Types of Artificial Intelligence
AI can be divided into different types on the basis of strength, breadth, and applications. On the basis of strength, AI can be divided into three ways, which are given below:
Weak or Narrow AI
Weak AI is AI that can be applied to specific domains. It can perform specific tasks intelligently but cannot be applied in other domains and learn new ones. It makes a decision based on the used algorithms and provided training data. Currently, we are at the stage of weak AI or Narrow AI.
Some examples are language translator, virtual assistants, recommendation systems, self-driving cars, etc.
2. Strong AI or Generalized AI
Strong AI can be applied to a wide variety of tasks that are independent of each other rather than for a specific task. It is able to learn new tasks to solve new problems by teaching itself new strategies.
It can perform a task at a human-level of intelligence.
3. Super AI or Conscious AI
It is also known as conscious AI. It means it has consciousness like humans and requires self-awareness. The creation of conscious AI is still a hypothetical concept, and creating such systems in the real world is currently very challenging and approx. To impossible task for the developers.
History of Artificial Intelligence
Nowadays, AI is evolving very rapidly, changing our world, and making our lives easier. But the word Artificial intelligence or AI is not new, as it was coined in the year 1956. It was started with various hypothetical situations, which have also been shown in Hollywood movies. So, let’s see the milestones and how it is evolved over various years till now:
||The first work of AI was done by Warren McCulloch and Walter pits in the year 1943. They proposed a model of artificial neurons, in which each neuron is divided as being “On” or “Off” with the switch.|
||In the year 1949, Donald Hebb provided an updating rule to improve the connection strength between the neurons.|
||In the year 1950, Alan Turing, the English Mathematician, published an article “Computing Machinery and Intelligence”. In this article, he proposed a test named the Turing test to demonstrate the intelligence of machines.|
||John McCarthy, who was an American computer scientist, first coined the term Artificial Intelligence at Dartmouth Conference.|
||The first chatbot was created in the year 1966 by Joseph Weizenbaum, which was named ELIZA.|
||The first intelligent robot, WABOT-1, was developed in Japan.|
||In this year, the first computer-controller autonomous vehicle was created, which was named Stanford Cart.|
||IBM Deep Blue defeated Gary Kasparov, the world Chess champion of that time, and became the first computer to defeat the world Champion in Chess.|
||AI started being used in the business world. Various big brands, such as Facebook, Twitter, etc., started using AI.|
||Various evolution has occurred in this duration, and day by day, new trends are occurring in the development of AI, such as Google Now, Project Debater, Duplex by Google, Spam filtering, deep learning, and many more.|
Applications of Artificial Intelligence (AI)
AI is evolving very fast and also impacting various fields such as finance, Marketing, Automobile, etc., to growing the business and making life easier. Let’s understand the various uses of AI in different fields.
- AI in Agriculture
Agriculture is the most crucial area for living life and requires various resources such as manpower, time, and money. So, to get significant results in agriculture production, AI can be used as agricultural robots for crop and waste management and predictive analysis to help the farmers.
- AI in Data Security
Data security is the most important aspect for each user on the internet, as there are so many cyber attackers active over the internet. AI can be used to make your data more secure and safe from the eye of cybercrimes. AEG bot, A12 platform, etc., are used to secure data and determine the software bugs in a better way.
- AI in Gaming
AI can be used in so many games available on the internet today. Like online Chess, cricket, and ludo, etc. The computer plays strategically to compete with human beings.
- AI in Entertainment (Netflix)
Netflix has developed its own movie recommendation for each of its users. These recommendations are made by Machine Learning. When you open your Netflix app and look at the type of movies that are recommended to you, they are quite different than others. This is because Netflix analyses each user’s personal details and tries to understand what each user is interested in, and which type of movie patterns each user has, then it recommends appropriate movies or series to them. This helps to engage a user on its platform by recommending them for watching the next movie and enhances the user experiences.
- AI in the Automobile sector
AI is used in automobile sectors also, and an example is self-driving cars. AI implements computer vision, image detection, deep learning while building self-driving cars, which can automatically detect any object or any obstacles and drive around without human intervention.
- AI in the Healthcare sector
Healthcare organizations use IBM AI (Watson) technology for medical diagnosis. IBM is one of the pioneers that developed AI software, especially for medicines. There are hundreds of healthcare organizations that use AI (IBM Watson) technology. In 2016, IBM Watson technology was able to cross-reference twenty million oncology records quickly and correctly diagnosed a rare Leukemia condition in a patient.
- AI in the Finance Sector
In the Finance sector, JPMorgan chase’s contract intelligent (COiN) platform uses AI, ML, and image recognition software to analyze legal documents in few seconds only. As per approximation, manually reviewing 12,000 agreements takes over 36,000 hours that is a lot of time-consuming. But now it is work of a few seconds after deployed it by AI machines. So, this is the main difference between AI and human work. However, AI can’t think and reason like humans, but it has better computational power than humans.