Reinaldo Padilha FranƧa, Ana Carolina Borges Monteiro, Rangel Arthur and Yuzo Iano
1.1 Introduction
Artificial Intelligence (AI) enables machines to learn from experiments, to adjust to new data inputs and to perform tasks as if they were human beings. It is a branch of computer science research that seeks, through computational symbols, to build mechanisms and/or devices that simulate the human beingās ability to think and solve problems, that is, to be intelligent. Nowadays, technologies ranging from computers becoming chess masters to autonomous cars rely on deep learning and natural language processing. With these technologies, computers can be trained to perform specific tasks by processing large amounts of data and recognizing patterns in that data [1, 2].
Early AI research in the 1950s explored issues such as problem-solving and symbolic methods, arriving in the 1960s, when the US Department of Defense became interested in this type of technology and began training computers to mimic basic human reasoning. Artificial Intelligence can be roughly defined as the ability of machines to think as human beings learn, perceive and decide which paths to follow rationally in certain situations [1, 2].
The desire to build machines capable of reproducing the human ability to think and act has existed for many years. This can be seen in the existence of autonomous machines. These early works paved the way for the automation and formal thinking that we see in todayās computers, including decision support systems and intelligent research systems that can be designed to complement and expand human capabilities [3, 4].
With the computational evolution, Artificial Intelligence has gained more strength; considering that its development has enabled a great advance in computational analysis, and the machine can even analyze and synthesize the human voice. At first, studies on AI sought only one way to reproduce the human capacity for thinking, but it was no different to all research that evolves [3, 4].
Nowadays there are technologies like āMachine Learningā which, instead of programming rules for a machine and waiting for the result, are able to let the machine learn these rules on its own from the data, reaching the result autonomously. Where we speak of āDeep Learningā, we are referring to a part of machine learning that uses complex algorithms to āmimic the neural network of the human brainā and learn an area of knowledge with little or no supervision. The system can learn how to defend against attacks on its own. There is also āNatural Language Processing,ā (NLP) which processing uses machine learning techniques to find patterns in large pure data sets and recognize the natural language. Thus, one example of applying NLP is sentiment analysis, where algorithms can look for patterns in social network posts to understand how customers feel about specific brands and products [5, 6].
AI automates repetitive learning and discovery from data, where it is different from hardware-driven robotic automation. Instead of automating manual tasks, AI performs frequent, bulky, computerized tasks reliably and without fatigue. For this type of automation, human interference is still essential in setting up the system and asking the right questions [7, 8].
AI adds intelligence to existing products, since in most cases Artificial Intelligence will not be sold as an individual application. Instead, the products people already use will be enhanced with AI functionality. Automation, chat platforms, robots and smart devices can be combined with large amounts of data to enhance many home and office technologies, from security intelligence to investment analysis [7, 8].
AI adapts through progressive learning algorithms to let data do the programming, finding structures and regularities in the data for the algorithm to acquire a capability: it becomes a classifier or a predictor. So just as the algorithm can teach itself how to play chess, it can teach itself which products to recommend next. And the models adapt when they receive more data. Retroactive propagation is an AI technique that allows the model to adjust through training and input of new data when the first response is not entirely correct [9, 10].
AI analyzes more data, and does so in greater depth, using neural networks that have many hidden layers; building a five-layer, hidden fraud detection system was almost impossible a few years ago. This has all changed with impressive computational power and big data, since it needs a lot of data to train deep learning models because they learn directly from the data. The more data that it is possible to put into them, the more accurate they become [9, 10].
AI achieves incredible accuracy through deep neural networks; examples ranging from interactions with Alexa to Google searches and Google Photos are all based on deep learning, where they keep getting more accurate as we use them. In the medical field, deep learning AI techniques, image classification and object recognition can now be used to find resonant cancers with the same precision as well-trained radiologists [9, 10].
AI gets the most out of data, when algorithms learn on their own, the data itself can become intellectual property. The answers are in the data; to be obtained is just need to apply AI to extract them. Since the role of data is more important than ever, it can create a competitive advantage. If these data already exist in a competitive industry, and yet everyone is putting similar techniques into practice, it is possible to get the one with the best data set [11, 12].
AI can be found today in healthcare providers, with Artificial Intelligence applications that result in medication and personalized x-ray readings. Personal assistants can act as coaches, reminding the user to take their medicine, to exercise or to eat healthy foods. In retail, AI provides features for online retailers, such as offering personalized recommendations and negotiating payments with consumers. Inventory management and site layout technologies are also enhanced with AI. In manufacturing, since AI can analyze IoT data from factories as it is transmitted from connected equipment to forecast load and demand using recurring networks, a specific type of deep learning network applied to sequential data. As in sports, where Artificial Intelligence is used to capture match images and provide coaches with reports on how to better organize the game, which includes position and strategy optimization on the pitch.
AI is a new wave of innovation, where economists call it the fourth industrial revolution, marked by the convergence of digital, physical and biological technologies, creating links between the boundaries of the three areas. Since AI is part of this next wave of innovation, it brings big changes in the way people and businesses relate to technology, share data and make decisions [11, 12].
The future of AI points to an increasingly transparent, ethically built technology that is part of everyday tasks, at work or in our personal lives, enhancing our cognitive abilities. AI can make humans more productive by releasing professionals from certain mechanical and repetitive tasks so that they can make the most of their ability to create and innovate in other industries [11, 12].
Therefore, this chapter aims to provide an updated review and overview of AI, addressing its evolution and fundamental concepts, showing its relationship as well as approaching its success, with a concise bibliographic background, categorizing and synthesizing the potential of technology.
1.2 Artificial Intelligence Fundamental Concepts
Artificial Intelligence is a branch of computer science that aims to develop devices that simulate the human ability to reason, perceive, make decisions and solve problems, in short, the ability to be intelligent; it is the ability of electronic devices to function in a way that resembles human thought, implying perceiving variables, making decisions and solving problems, operating in a logic that refers to reasoning [1, 13].
According to the definition, intelligence is the āfaculty of understanding, thinking, reasoning and interpretingā, that is, the āset of mental functions that facilitate the understanding of things and factsā, containing the ability to take advantage of the effectiveness of a situation and use it in the practice of another activity, as well as the ability to resolve new situations quickly and successfully, adapting to them through the acquired knowledge; as well as āartificialā, it was āproduced by man and not by natural causesā [3, 14].
Therefore, Artificial Intelligence is a field of science, whose purpose is the study, development and the use of machines to carry out human activities in an autonomous way. It is developed...
