Artificial intelligence (AI) is a branch in computer science that deals with machine intelligence. It is the study and design of intelligent agents (a system that recognizes the environment that it is in and uses all available strategies to ensure its success). It can also be defined as “the science and engineering of making intelligent machines”. Over many years, machines have been found to be intelligent just like human minds. They can carry out processes that are even beyond the human mind. The study of artificial intelligence was founded under claims that the intelligence of human beings could be stimulated by machines. AI is further subdivided into non-communicative subfields.
History of Artificial Intelligence
The likeness to human beings which is believed to have human intelligence was built in almost all civilized nations. In Egypt and Greece, some of these likenesses were actually worshipped. By the 20th century, stories of the energetic beings were all over and their fortunes used to converse the fears, hopes and moral anxieties. Mathematicians and philosophers came up with habitual reasoning to study the notion behind AI. This led to the invention of “(programmable) digital- electronic computer”. In the assumption of computation, Turing recognized that the symbols “0” and “1” could help in making mathematical judgments. This invention motivated other researches who saw the likelihood of establishing “an electronic brain” (Nilsson 1998).
In 1956, a conference was held in Dartmouth College in which the study of AI research was founded. Some of the attendees of the conference became great leaders in AI research for many years. Students started writing programs and computers were found to solve algebra problems, provided logical theories and were said to speak in English (Cerri & Dochev 2000). By 1960, researches were all over the world especially in U.S which set aside laboratories to facilitate research. The founders of AI were optimistic that the new field would eventually thrive. Some went ahead to predict what would happen in future. According to them machines would be capable of doing all kinds of work that was done by man in a period of twenty years. However, they did not consider some of the difficulties they were going through which led to the collapse of the field in 1974. It was revived in 1980 by some commercial expert systems and five years later AI market was over a billion dollars.
In the late 20th century and early 21st century, great achievements were evident in the AI field. Today, AI is used for medical diagnosis, logistics and data mining. The success of the artificial intelligence can be attributed to several factors: the implausible computer power, formation of close ties between AI and other fields which were working to solve the same problems and the dedication from the AI researchers (Cerri & Dochev 2000).
One of the commonly known AI agents is the internet. For quite some time, the AI systems have been common in web-based applications. Everybody is relying on the internet for the day to day operations. AI technologies rely on the use of internet tools such as website aggregators, recommender systems, and search engines. The use of the internet has made it possible for the rearrangement of the previously isolated fields that needed to be reorganized. It was realized that the sensory systems such as speech recognition and vision could not be relied on for the delivery of information related to the environment. This could only be possible through proper planning and reasoning.
One reason behind the creation of a complete intelligent agent was to facilitate the close contact between AI and other subfields such as economics and control theory that needed the use of agents (Cerri & Dochev, 2000).
The early founders of AI have not been contented with the AI’s progress. According to them, AI should not be concentrated on the creation of improvement on applications that are used in specific task such as recognizing speech or driving. Instead, focus should be directed to AI’s roots of striving to create machines that can learn, think, and be able to create.
There is one major problem encountered in creation of intelligence. This can be broken down into precise sub-problems. These problems are the specific qualities that most researchers would like to be exhibited by the intelligent system. Let’s discuss these qualities:
Logic and Reasoning
The early founders of AI established algorithms that emulated human reasoning when solving mysteries or when making logical assumptions. By early 1990s, AI researches had discovered methods for dealing with doubtful information and were able to employ concepts that could be used in economics. In order to solve some of the complex problems, algorithms require use of vast computational resources and more computer time. The amount of computer time or its memory gets exorbitant as the problem exceed a certain size. Among the priorities of AI research is the search for more proficient problem solving algorithms (Nilsson, 1998). Today, human beings use their ruling in solving problems rather than depending on the assumption methods that were replicated by early AI researchers.
Some the key components of AI research are knowledge engineering and representation. Most of the troubles that are anticipated to be solved by machines necessitate the use of a wide range of facts about the world at large. This is because these machines need to represent properties, objects, events, relationships, effects and causes among others. Some of the problems related to knowledge representation in AI research are qualification problem and commonsense knowledge. The human mind works under assumptions which AI researchers referred to as the qualification problem. In every rule there are many exception thus assuming everything can bring some controversies. Consequently, AI has been researching on how to end this predicament. Some of these solutions require huge amounts of ontological engineering. They are aimed at facilitating computer understanding of many concepts that will enable it to learn from other sources for example the internet. One in the search for these solutions is that human knowledge cannot be represented as statements and only occurs in the non-conscious mind.
Another sub-problem in AI research is planning. Goals and objectives have to be set by intelligent agents which are also responsible for preparing strategies of achieving them. These agents have to understand the current state of the world and make predictions about the future of their actions. Choices have to be made that utilizes the available resources to the maximum. According to Russell & Norvig (2009), planning comes before scheduling, that is, “we divide the overall problem into a planning phase in which actions are selected, with some ordering constraints, to meet the goals of the problem, and a later scheduling phase, in which temporal information is added to the plan to ensure that it meets resource and deadline constraints”p401
Since the foundation of AI research, machine learning has been a key element. Learning can be divided into two unsupervised and supervised. Supervised leaning require the use of arithmetical regression and arrangement. Arithmetical regression is an effort of finding out functions that can fabricate output from specified inputs using arithmetical outputs or inputs. In arrangement, categories are differentiated from each other. Russell & Norvig (2009), states that “in unsupervised learning the agent learns patterns in the input even though no explicit feedback is supplied. The most common unsupervised learning task is clustering: detecting potentially useful clusters of input examples” p695.
The intelligent agent gets rewards for any good responses evidenced and receives punishments if bad responses are discovered. There is an area in the studies of computer science recognized as computational learning assumption that is accountable for facilitating machine knowledge of algorithms. Machines are able to read and comprehend human verbal communication through a procedure called natural language. Applications used in this process are machine translation and retrieval of information (Chrisley & Begeer, 2000). It is believed that, through reading a text that has been fed to the computer, machines can be able to understand the human language.
Decision tree induction is one of the easiest and effective methods of learning in machines. This method involves description of the representation, hypothesis space, and ways of learning good hypothesizes.
Creativity and Social Intelligence
Creativity refers to the theoretical and practical ability of generating inputs from output. This is studied in a field in AI research called artificial imagination. Intelligent agents require both social and emotional skills to be able to carry out their work. By considering the poignant state of others, an agent can be able to forecast their future actions. Social intelligence requires the use of decision theory and game theory to make the predictions (Chrisley & Begeer, 2000). A computer is termed as having good human interactions if it’s able to display emotions. It should be able to have ordinary sentiments and be responsive to the human relations.
Artificial Intelligence (AI) is the study of machine intelligence. Machines have the capability of performing some of the functions that are carried out by the human mind. This study has evolved over along period of time and advancements have been made to it. Today, AI is used for medical diagnosis, logistics, and data mining. The success of the artificial intelligence can be attributed to several factors: the implausible computer power, formation of close ties between AI and other fields which were working to solve the same problems and the dedication from the AI researchers. AI research requires the use of the internet. It uses web-based applications and internet tools such as search engines and web aggregators.
Problems in AI research can be split into various sub-problems. They include: deduction and reasoning, knowledge representation, planning, learning, and social intelligence, and creativity. The amount of computer time or memory gets exorbitant as the problem exceeds a certain size. Goals and objectives have to be set by intelligent agents which are also responsible for preparing strategies of achieving them. These agents have to understand the current state of the world and make predictions about the future of their actions.
Cerri, S. A. & Dochev, D. (2000). Artificial intelligence: methodology, systems, and applications: 9th international conference, AIMSA 2000 Verna, Bulgaria, September 2000: proceedings. Volume 1904 of Lecture Notes in Computer Science. Springer
Chrisley, R. & Begeer, S. (2000). Artificial intelligence: critical concepts, Volume 4 Artificial Intelligence: Critical Concepts. New York: Taylor & Francis
Nilsson, N. J. (1998). Artificial Intelligence: a new synthesis. The Morgan Kaufmann Series in Artificial Intelligence Series. Morgan Kaufmann
Russell, S. J. & Norvig, P. (2009). Artificial Intelligence: A Modern Approach. Prentice Hall series in artificial intelligence. New York: Prentice Hall