Areas of application of artificial intelligence systems
In the 20th century, artificial intelligence was one of the main themes of science fiction stories and books. In 2020, people no longer notice how they use AI systems in their daily lives. Whether they are engaged in reading mail or watching films in Kinopoisk. Artificial intelligence systems range from voice assistants to medicine and space exploration.
Introduction
Artificial intelligence (AI) is a software package that is able to reproduce human skills: plan, solve problems, give advice, as well as learn and improve their work in the process of completing tasks. For example, music recommendation services that study your preferences and "throw" similar music can be called a kind of artificial intelligence.
Human thinking is based on neurons in the brain, while AI thinking is based on neural networks. They enable systems to acquire new skills in much the same way that humans do.
How to distinguish artificial intelligence from a simple algorithm for collecting and issuing data? The main feature of AI and its difference is the ability to learn and improve in the process. That is, the more you use the technology, the better it adjusts to your needs, while a conventional system continues to execute only the originally specified algorithm.
Why create artificial intelligence systems
The use of artificial intelligence in work allows not only to automate any process, but also to customize it in accordance with the specific task of a person, department or production. The work of AI becomes more efficient over time due to constant learning - the more the neural network knows the details and needs, the better it functions.
Initially, many assumed that AI would only be able to translate texts, recognize objects and capture the meaning of human speech. But by 2020, the list of skills has expanded so much that listing it would take more than one page. Today's AI is present in many areas of our life - on the Internet, medicine, business and even transport.
Artificial Intelligence on the Internet
The introduction of artificial intelligence on the Internet began back in 2011. At that time, work was started on the Google Brain project. The result was the introduction of new AI-controlled options into the Google search engine:
- Google translator with direct translation - just click on the microphone icon to activate the voice input function. The system recognizes speech and quickly translates words or phrases into the desired language.
- Voice assistant Google Assistant - you can start searching for information, give assignments, plan things.
- The Google Photo algorithm has learned to recognize texts and objects in images. Thanks to this function, you can easily find photos by describing objects or people that are present in the image.
Additionally, a new search algorithm Google BERT was introduced . Now the search results are formed based on the analysis of not only keywords, but also whole sentences. For this, a bidirectional neural network-encoder is used, which for the user improves the relevance of the pages displayed in the output. Thanks to this technology, getting the desired result in the SERP becomes much easier.
The creators of the Yandex search engine also did not stand aside and carried out work on the introduction of AI. Part of the tasks that AI performs - the system is used in the voice assistant "Alice", search algorithms for speech and image recognition, getting weather reports.
Artificial intelligence in transport and logistics
Since the 2000s, many automakers have taken up the development of self-driving cars. Innovators include Nissan, BMW, Honda, Volkswagen and Audi. Unmanned vehicles are based on radar, light and distance detector, GPS and special cameras. All incoming data is analyzed by AI to make decisions on the road.
Amazon uses unmanned aerial vehicles to deliver goods. The first parcel sent in this way arrived at the recipient at the end of 2016.
A smart road system has been created in Yekaterinburg based on artificial intelligence. It is a mechanism for regulating traffic flows, unloading roads and ensuring the smooth movement of ground passenger transport. At the same time, the system records violations and sends fines. Information comes from vehicle detectors, photo and video recording complexes, on-board equipment and other devices.
Artificial intelligence in finance
International payment system MasterCard has introduced an additional service Decision Intelligence. The company notes that losses from security errors exceed losses from fraud. Implementation of Decision Intelligence has increased the level of the company's profit.
The main functions of this technology are to improve the accuracy of confirmation of financial transactions and reduce the likelihood of false rejections when transferring funds. The system works on the basis of a neural network. When analyzing financial transactions, large amounts of data are processed from sources connected to the system. Numerous factors are taken into account, including the type of customer purchase, location, and time of day. Thus, the number of false positives of the built-in security system is minimized, "normal" transactions go through without problems.
PayPal also uses AI to detect suspicious activity. The system analyzes transactions according to several behaviors developed by the electronic system. This reduces the number of fraudulent transactions and false alarms.
Artificial intelligence in lending services makes it easier to analyze the history of borrowers, accelerates decision making on granting loans and reduces the number of late or unpaid payments.
Artificial intelligence in business and trade
Artificial neural networks are actively used in retail and business. The most widespread use of AI was found by Walmart, which owns a large retail chain. With the help of neural networks, it was possible to automate the payment system, simplify the accounting of goods and ensure prompt delivery by drones.
Since 2017, Bossa Nova robots have been operating in the chain's stores . They inspect all supermarket branches three times a day, checking store shelves for the entire assortment, incorrect price tags or purchased products. The collection and analysis of information takes place thanks to specially developed software with elements of artificial intelligence. The data is transferred to the logistics department.
The peculiarity of robots is that they are not only equipped with sensors for scanning goods, but also have a special security system. She is responsible for recognizing nearby objects. This prevents the Bossa Nova from colliding with carts, people, shelves and product groups.
At Lowes electronics stores, LoweBot robots act as additional consultants . They navigate the sales area, helping customers find the right products. Robots ask customers simple questions to figure out which technique is right for them. In the future, the AI remembers the buying activity and more accurately understands which products should be offered first. Also LoweBot talk about the current discounts. In addition, robots monitor the assortment of products so that store employees can timely lay out missing goods.
Artificial intelligence in medicine
Artificial intelligence recognizes pathologies on X-rays, mammography, MRI, CT. With the help of AI, doctors identify lung diseases, Alzheimer's disease. It takes a few seconds for artificial intelligence to study the research results, and a doctor - much more, for example, 20-30 minutes.
The most famous computer diagnostician is IBM Watson . Millions of medical documents and medical records are loaded into his memory. It is noteworthy that IBM Watson makes diagnoses more accurate than doctors. But the immediate attending specialist is only guided by the advice of AI, independently deciding on the diagnosis and methods of treatment.
Another example of Russian AI is Botkin.AI, created to help oncologists. The system conducts a multilateral analysis of examination results, gives tips and advice to doctors.
Artificial intelligence is also used to accelerate the recovery of athletes. Based on the complex of factors, AI selects the most effective treatment complex.
Artificial intelligence in defense and military affairs
AI is also being introduced into the military-industrial sphere. In 2018, it became known that the US Army is developing a system that will be able to recognize people's faces in the dark and through walls using a thermal imager. Its main purpose will be to identify gang leaders in locations where hostilities are taking place.
Another AI called ALPHA is being actively tested. It is designed to control unmanned fighters and participate in military air operations. Testing the ALPHA on simulators resulted in the computer defeating two human opponents while flying four fighters simultaneously.
Artificial intelligence in security systems
AI is used in security systems primarily for face recognition and personal identification. Additionally, "smart" systems are used to detect dangerous objects and substances.
In addition, AI provides assistance in cyber security. By analyzing data sets of threats, artificial intelligence improves security response times and empowers them to respond more quickly. For example, IBM offers a range of solutions for cybersecurity professionals. Watson technology searches for the relationship between threats and provides actionable recommendations. As a result, you can respond to threats faster and more confidently.
Artificial intelligence in space systems
One example is NASA's Curiosity robot, designed to study the composition of Martian soils and atmospheric components. Thanks to the presence of AI, Curiosity can not only study the terrain, but also memorize safe paths, as well as create new routes, taking into account previously obtained knowledge about the nature of the soil or soil.
Another robot powered by artificial intelligence is Lauron . It was developed at the Karlsruhe Institute of Technology. This pedestrian robot has been designed for statically stable walking on uneven terrain. Thanks to a flexible control system, it adapts to different landscapes. The peculiarity of the robot is the presence of six legs with special hooks. Lauron is used to explore areas in space objects.
The machine collects information about the environment and autonomously plans the path to the goal. While moving, Lauron “sees” obstacles, and then either passes over them, or bypasses them if the obstacles are too high.
The pedestrian robot is designed to inspect and maintain areas that are difficult and dangerous for humans. So, among his tasks is the study of the surfaces of volcanoes and other planets.
Artificial intelligence in sports
The organizers of the baseball, football and basketball teams analyze the individual data of the players, their technique, and physical condition. Artificial intelligence is using this data to help predict the potential of athletes.
Another example of using AI technology is predicting match results. Special systems have been created by UBS, Commerzbank and Microsoft. Many factors are taken into account in the calculations, for example, the experience and physical condition of the players, weather conditions, the location of the meetings. All this is used to make sports predictions.
Artificial intelligence in the system of municipal government
The introduction of AI into municipal management is intended to make it more effective, to influence the audience correctly, increasing the chances of getting the desired result. Barack Obama, in the second presidential election, hired a team of professionals that used AI. Neural networks were responsible for calculating the best days and choosing locations for campaigning events, which gave an advantage of 10-12%.
Artificial intelligence in culture
In October 2018, the first painting was sold, on which artificial intelligence worked. It was purchased for $ 432,500. When creating the work, a special generative adversarial network algorithm was used, which analyzed more than 15,000 portraits of artists of the 15th-20th centuries.
In the music field, Warner Music recording studio has signed a long-term contract with the robo-performer Endel . A total of 20 albums will be released. The peculiarity of Endel is that it is based on artificial intelligence with a special algorithm.
The neural network is capable of not only writing ordinary music, but also creating individual compositions depending on the listener's mood. In the latter case, a person's personal data, his health, location and other information are analyzed. Based on this, a unique track is created for a specific person that can improve mood and reduce feelings of anxiety.
Artificial intelligence in education
Due to the introduction of AI in the future, the education system will develop in two directions. The first one is adaptive. Its main task is to solve the problem of different academic performance among students. AI will analyze the results of students and, on their basis, adapt the order of courses, additionally informing teachers about the degree of mastery of the material.
The second area is proctoring. The aim is to ensure that students are supervised during tests and exams. The system keeps track of whether schoolchildren or students are talking to each other, how often they avert their eyes from a notebook or computer, and whether they use pocket gadgets. When violations are detected, the AI immediately sends an alert to the proctor - a specialist responsible for monitoring the passage of tests.
Artificial intelligence in the judicial system
China was among the first to use AI. Neural networks are still used as assistants. They analyze large amounts of data from government storages, take into account the characteristics of a particular person, and then make a decision on his guilt or innocence.
Some machines based on artificial intelligence are able to predict future human offenses based on statistical information. Thus, they plan to reduce the overall crime rate.
Artificial Intelligence in Agriculture and Livestock
Companies in the agrarian industry such as Agworld, Farmlogs, Cropx, AGCO are actively creating and implementing AI systems in various areas of agriculture and animal husbandry. For example, drones with radars and GPS monitoring are trained and then used to deliver hazardous chemicals and spray crops.
CNH Industrial manufactures self-driving tractors. They perform the same tasks as conventional agricultural machinery. But productivity is higher due to the ability to work continuously.
AI machines are capable of spraying, weeding, monitoring livestock in a herd, and other labor-intensive agricultural tasks. In animal husbandry, artificial intelligence systems are used to detect diseases in livestock, to identify good breeding animals.
Artificial intelligence for military purposes
To preserve sovereignty, many states are actively creating drones and drones. They can be combined into a controlled swarm for massive attacks. Coping with the many drones will be nearly impossible. There is little data on the AI being introduced into the military sphere - the information is secret, it is not in the public domain.
Artificial intelligence in industry
In the industrial industry, the introduction of AI technologies is associated with the automation of production processes and the reduction of the staff. A machine with an artificial intelligence system assembles parts, easily memorizing the sequence of actions and the correctness of fastening the elements, accurately calculates the data and optimizes the assembly: in what order to arrange the parts, how to secure them more securely.
Korean company LG has planned to open a fully automated plant in 2033. All processes will be performed using artificial intelligence - from the purchase of raw materials and consumables to the manufacture of products and their shipment. Additionally, the AI will be responsible for monitoring the wear and tear of production equipment, pricing, fulfillment of plans, etc.
Application of the artificial intelligence system as an expert system
Smart machines have knowledge in a specific area and can act as specialists, making analysis and proposing solutions for the tasks at hand.
For example, there are expert systems for diagnosing diseases, making financial projections, or preparing the shortest routes for transport in logistics systems.
Some expert systems can train people in a specific field, using the inherent knowledge for this. Application areas are:
- Chemistry;
- agronomy;
- geology;
- engineering;
- meteorology;
- military science.
Problems of using artificial intelligence
When implementing AI systems, specialists face a number of difficulties:
- Smart technologies are available only to large companies or ministries that can afford to invest large sums and get a return on it only after a few years. So, for the development of the Digital Economy program, which includes technologies with artificial intelligence, the Ministry of Telecom and Mass Communications put 1.118 trillion. rubles .
- After purchasing robotics or "smart" programs, the stage of their integration into the business begins, where there are problems with personnel. In 2020, there is a shortage of specialists capable of performing all the necessary AI add-ons for their introduction into business.
- Until now, it remains unclear who should be responsible for emerging incidents when working with AI - the creators or those who use it.
- Most AIs require massive amounts of information to train and run successfully. At the same time, there is a problem of data confidentiality. Created AI projects are not always provided with the necessary level of security for data safety.
The near future of artificial intelligence
Trevor Sands, an AI researcher at Lockheed Martin, makes optimistic predictions . He believes that in the next 5-15 years there will be several breakthroughs in this area. Sands does not exclude that used on most of the innovative solutions will come from the corporation Google.
However, Trevor Sands claims that it is unlikely that it will ever be possible to create an AI that reaches the human level. The reason is that the human brain is endowed with great plasticity and more than 86 billion neurons. It is easy for him to engage in various fields of activity, to quickly switch between them. Whereas AI is provided with only a few thousand or a million artificial neurons. Even if active work is carried out to improve technology, computers with AI will still not be able to compete with the human mind.
Conclusion
Artificial intelligence is used in at least 16 areas. Its use made it possible to simplify and improve the quality of tasks performed, reduce the load on human resources and optimize many processes.
In medicine, the speed of making diagnoses has increased and their accuracy has increased. In the educational sphere, the quality of education has improved due to the fact that the individual characteristics of students have begun to be taken into account. The transport sector has been replenished with unmanned vehicles. In logistics, the accounting of goods has been simplified, the number of errors has decreased. The arrival of AI in agriculture has simplified fertilization and processing of crops, increased productivity in harvesting.
The benefits of artificial intelligence are difficult to assess objectively, but in these areas, it really simplifies many processes. As AI is constantly improving and learning, predictions remain optimistic.
No comments: