Artificial intelligence and its powerful capabilities
The rapid growth and powerful capabilities of artificial intelligence (AI), and with the transformation brought about by artificial intelligence in various industries, have made business leaders and the general public believe that we are close to the peak of artificial intelligence research and the realization of the maximum potential of artificial intelligence. However, understanding the types of AI that currently exist provides a clearer picture of the available AI capabilities and the long road ahead in AI research. Artificial intelligence has different categories that we will get to know.
What is artificial intelligence?
The many applications and prevalence of artificial intelligence need no introduction. Today, artificial intelligence is not just a “buzzword,” but a reality that has become practical in all aspects of our lives. Intelligent machine companies pursuing different purposes and uses have revolutionized the industry.
Artificial intelligence is the process of creating intelligent machines using big and massive data. Machines learn from past experiences and perform human-like activities. Due to these machine activities, accuracy, speed and efficiency are improved.
Artificial intelligence uses complex patterns and methods so that machines can make decisions independently and on their own. Machine learning and deep learning are the two basic components of artificial intelligence. Currently, artificial intelligence is used in almost all business sectors.
Types of artificial intelligence
Basically, AI types can be classified into two aspects, including AI types based on job-based capabilities and AI types. In this section, a list of each of these different aspects of AI will be given first, and then each will be described separately.
Types of AI based on capabilities:
Narrow AI
General AI
Super AI
Types of Job Based AI:
“Reactive Machines”
Limited Theory
Theory of Mind
Self-awareness
The types and types of artificial intelligence are mentioned from two aspects, and they are also shown graphically in the image below.
Narrow AI
Finite AI, also known as “weak AI”, only focuses on a specific job and can no longer act on it. Finite AI focuses on one subset of jobs. Applications of this type of AI are becoming more common in our lives and with the development of Machine learning and deep learning are more prevalent.
Apple’s intelligent voice assistant Siri is an example of limited AI that works with a set of predefined functions. Out-of-band Apple Siri activity can disable it.
The IBM Watson assistant is another example of limited AI that processes information and answers user questions using computing, machine learning, and natural language processing. Once upon a time, the IBM Watson artificial intelligence was able to defeat its human competitor in the popular game program “Jeopardy” and become the champion.
Some other examples of limited AI include Google Translate, image recognition software, and Google’s search engine optimization algorithms.
General AI
Artificial general intelligence, also known as Strong AI, is another type of artificial intelligence that can understand and learn any intelligent human function. Artificial general intelligence enables machines to apply skills and use science in different fields.
Artificial intelligence researchers have not yet succeeded in achieving strong artificial intelligence. A necessary condition for achieving this goal and for machines to be fully intelligent is to program a full range of cognitive capabilities. Microsoft has invested $1 billion in the open AI Institute to develop artificial general intelligence.
Fujitsu has built one of the fastest supercomputers in the world, the K-computer. This project is considered one of the most important activities in the direction of achieving powerful artificial intelligence. A one-second simulation of human nerve function takes about 40 minutes. Therefore, it is not easy to say whether the realization of strong artificial intelligence will be possible soon.
Fujitsu’s K supercomputer
Tianhe-2 is a supercomputer developed by the National University of Defense Technology in China. This supercomputer has the capacity to process 86.33 petaflops (quadrillion) arithmetic operations per second (cps). Although this sounds exciting, the human brain is estimated to perform one exaflop, that is, a billion computational operations per second.
The Tianhe-2 supercomputer is made in China
Types of AI based on job
As mentioned earlier, to describe the types of AI systems, AI can be classified based on how it works. Therefore, in this section, the types of AI are discussed in functional terms, which generally include 4 categories. First, interactive machines and then 3 other types of artificial intelligence were introduced from this side.
“Reactive Machines”
An interactive machine, also known as a passive or passive machine, is an early form of artificial intelligence that does not use past experiences and records for prediction and uses only current data. The interactive machine understands the surrounding world and interacts with it. Certain functions are specified in the interactive machines and the possibility of working thereafter is not possible for this type of machines.
The artificial intelligence Deep Blue (owned by IBM), which defeated the great chess master Garry Kasparov, is an interactive machine that reacts to the location of pieces on the chessboard. Deep Blue cannot take into account its past experience and improve itself by learning from practice
previous t. This machine simply recognizes chess pieces, knows how they move and predicts what the next moves for him and his opponent will be. Deep Blue does not take any of the data into account prior to the present moment and, by looking at the chessboard, selects and performs a move among the possible moves.
Limited Theory
Finite memory systems are a class of artificial intelligence that learns from past data to make decisions. Memory in such systems is short term. These systems are only allowed to use the data in a certain period of time and are not allowed to archive the data in their experiment library. This type of technology is used in self-driving cars.
Here is the limited memory operation of AI in self-driving cars.
Limited theory AI monitors how vehicles move in the surrounding environment, in real time and over time.
This type of machine, in addition to receiving a continuous data stream, stores static data including road signs and traffic lights.
This data is called up when the vehicle decides to change course; So as not to disrupt the movement of other drivers or not to cause an accident.
Mitsubishi Electric is conducting research to improve this type of AI in various use cases, including self-driving cars.
Theory of Mind
This type of AI is an advanced class of technology and is currently only a concept. A theory of mind is a type of function-based AI, which states that people and things in the environment can change feelings and behaviors. This kind of AI must understand emotions, feelings, and thoughts. Although this field has made a lot of progress, it is not yet developed.
A practical example of a theory of mind is the Kismet machine, which consists of a robot in the shape of a human head created by a researcher at the Massachusetts Institute of Technology (MIT) in the late 1990s. Kismet can mimic and recognize human emotions. Both of these abilities are important advances in theory of mind and artificial intelligence, but Kismet is unable to make eye contact and pay attention to humans.
Sophia from Hanson Robotics is another example of a theory of mind. Cameras placed in Sophia’s eyes give her a sense of sight using computer algorithms. Sophia can recognize people, track gazes, and make eye contact.
Self-awareness
This kind of AI is limited to the premise. These systems understand their internal characteristics, states and conditions as well as human emotions. These machines will be smarter than the human mind. The conscious system is not only able to understand and evoke emotion in interaction with others, it also has its own emotions, needs, and beliefs.
Types of applications of artificial intelligence
Here are 14 notable applications of AI, some of which we have briefly described below:
Application of artificial intelligence in e-commerce
Personal shopping
AI-equipped assistants
Fraud prevention
Application of artificial intelligence in education
Automating the administrative activities of teachers
Intelligent content production
Voice assistants
personalized learning
Application of artificial intelligence in lifestyle
Application of artificial intelligence in navigation
Application of artificial intelligence in robotics
Application of artificial intelligence in human resources
Application of artificial intelligence in medicine and healthcare
Application of artificial intelligence in agriculture
Application of artificial intelligence in games
Application of artificial intelligence in the car
Application in social media
Application of artificial intelligence in marketing
in chatbots
Applications of artificial intelligence in finance
Application of types of artificial intelligence in life
Artificial intelligence has had a huge impact on our way of life. In this section, we have described some of these applications.
Self-driving cars
Spam filter
Face recognition
Recommendation system
The use of artificial intelligence in the self-driving car industry
Car companies like Toyota, Audi, Volvo, and Tesla are using machine learning to train the computers in their cars. Artificial intelligence in self-driving cars, like humans, recognizes objects in the environment, so that accidents do not occur while driving.
Application of artificial intelligence in building spam filtering systems
Every day, spam emails sent through spam or trash folders are filtered and refined by artificial intelligence systems, and we make use of them regardless of these intelligent systems. Gmail managed to achieve a filtering capacity of approximately 99.9%.
Application of types of artificial intelligence in face recognition
Our favorite devices such as phones, laptops, and PCs use facial recognition methods and facial filters to enable secure recognition and access. In addition to personal use, facial recognition is widely used in various industries and in the aforementioned security cases.
Application of types of artificial intelligence in recommendation systems
Many of the platforms we use in our daily lives, such as e-commerce, entertainment websites, social media, video sharing platforms such as YouTube, etc., all use “recommender systems” to receive user information, and provide personalized recommendations to users. They use it to increase interaction. This application is widespread and frequently used in various industries.
Application of types of artificial intelligence in medicine and healthcare
Artificial intelligence has various applications in medicine and healthcare. Artificial intelligence is being used in healthcare to build advanced machines that can diagnose diseases and identify cancer cells. using Albia
Laboratory data and other medical data, AI can help analyze chronic conditions to ensure early diagnosis. AI can discover new drugs using a combination of historical data and medical intelligence.
Types of AI algorithms
Many algorithms are used in machine learning and data mining for classification, clustering, and data modeling, and some of the most popular and widely used algorithms are presented in this section. Types of AI algorithms can be grouped into three main categories: “supervised learning,” “unsupervised learning,” and “deep learning.”
Types of supervised learning algorithms in artificial intelligence
Supervised learning is a subset of machine learning and artificial intelligence defined by using labeled datasets to train models that accurately classify data or predict outcomes. Supervised learning helps organizations solve a wide variety of real-world problems. The most important supervised algorithms are listed below.
- Linear Regression
- Logistic Regression
- Artificial Neural Network (ANN)
- Support Vector Machine | SVM
- “K-Nearest Neighbor” (K-Nearest Neighbors | KNN)
- Types of supervised learning algorithms in artificial intelligence
Unsupervised learning uses machine learning algorithms to analyze and aggregate unlabeled data sets. These algorithms discover hidden patterns or clusters hidden in the data without the need for human intervention. The method’s ability to detect similarities and differences in information makes it an ideal solution for exploratory data analysis, client clustering, and image recognition. Here is a list of the most important unsupervised learning algorithms.
- clustering
- (K-means Lloyd| )
- Mattress Series (Hierarchical)
- Association Rules
- Reinforcement learning in artificial intelligence
Reinforcement learning is a machine learning technique that relies on rewarding desirable behaviors and/or punishing unwanted behaviors. In general, a reinforcement learning agent is able to understand and interpret his environment, take action, and learn through trial and error.
Computational intelligence
Computational Intelligence (CI) is the theory, design, application, and development of biologically and linguistically motivated computing models. Traditionally, the three main pillars of CI include neural networks, fuzzy systems, and evolutionary computing. Of course, many computational models inspired by nature have evolved over time. Therefore, CI is an evolving field that, in addition to the three main parts, computational models such as environmental intelligence, artificial life, cultural learning, “Artificial Endocrine Networks”, social reasoning and “Artificial Hormone Networks” (Include networks of synthetic hormones.
Computational intelligence plays an important role in the development of successful intelligent systems, including games and recognition development systems. In the past few years, there has been extensive research in the field of deep learning, especially deep convolutional neural networks. Today, deep learning has become the main method in artificial intelligence. In fact, some of the most successful AI systems are based on CI. In the following, some of the most important algorithms and methods used in computational intelligence are presented and described.
neural networks
Artificial Neural Networks (ANNs) are distributed parallel networks, inspired by the human brain, that have the ability to learn and generalize from examples. This research field includes feed-forward NNs, recurrent NNs, self-organizing NNs, deep learning, convolutional neural networks, etc.
Frequently Asked Questions
In this section, a number of common and frequently asked questions related to the types of artificial intelligence and related concepts have been addressed in order to resolve ambiguities in this regard as much as possible.
What is the most common type of artificial intelligence?
The most common type of artificial intelligence is limited artificial intelligence. Limited AI cannot operate outside its scope, as it is only trained for a specific task. This type of AI represents all types of AI that exist, including the most complex and capable AI created to date. Limited AI actually refers to AI systems that can only perform a specific task independently and using human-like capabilities. These machines can do no more than they are programmed to do, and so have a very limited or narrow range of capabilities.
What is strong and weak artificial intelligence?
Weak AI is an AI that focuses on a limited task or problem. All current AI systems are actually weak AI. A strong artificial intelligence is a machine that has consciousness and feelings. Strong AI is a virtual machine that is able to think and do things on its own like a human.
What is the difference between machine learning and artificial intelligence?
Artificial intelligence refers to systems or machines that mimic human intelligence to perform tasks and can improve themselves iteratively based on the information it collects. Machine learning, a subset of artificial intelligence, uses algorithms to automatically learn ideas and patterns and use that learning to make better decisions.
In which areas is artificial intelligence used the most?
The use of artificial intelligence is more superior
in the following fields and areas:
- Retail, shopping and fashion.
- Security and surveillance.
- Mathematical analysis.
- manufacturing.
- Inventory Management.
- Self-driving cars.
- Healthcare and medical imaging analysis.
- Warehouse supply chain and logistics.
conclusion
In this article, the types of artificial intelligence are discussed. Artificial intelligence is divided on the basis of its capabilities and functions. Types of ability-based AI include the following.
- Limited artificial intelligence
- Strong artificial intelligence
- superior artificial intelligence
Also, the types of AI are divided into the following four categories based on their function.
- negative machines
- limited theory
- theory of mind
- self conscious
In this article, in addition to the above, topics such as types of applications, especially applications of AI in lifestyle and medicine, as well as types of AI algorithms, have been covered. We may be far from self-aware machines that can solve all problems, but we should focus our attention on understanding how machines learn and train independently. We need to improve the machine’s ability to make decisions based on past experiences.
