Artificial Intelligence and Expert Systems (Essay Sample)

Artificial Intelligence and Expert Systems Maytte Sanchez - 50109065 Department of Computer Science, Texas A&M University – Commerce CSCI 526 - Database Systems Dr. Lee December 4, 2022 Abstract Databases are collections of data that have been systematically arranged and kept, most often in digital form. Data organized in a way that makes it simple to retrieve, organize, and alter is called a database. From a technological point of view, the paper highlights the most foreseeable benefits and drawbacks of AI and expert systems in relation to database utilization. Throughout this paper, this history of artificial intelligence and its evolving databases will be discussed in relation to the possibilities that artificial intelligence holds. While once used as a simple “robot” that would perform specific tasks being given by a computer, AI would evolve into something much greater, much more beneficial to society, or possibly to the actual detriment of society itself. In order to properly discuss artificial intelligence, its background will be detailed along with its modern day issues and topics that its databases relate to, including but not limited to its economic impact, the overwhelming costs needed to support it, and its role in privacy when it comes to our everyday lives. This also includes how we can create a relational database using artificial intelligence, for example having a chat bot to submit technical tickets. After discussing artificial intelligence’s history, its evolution, and the possibilities that lie ahead, the statement made by Elon Musk above may not seem so delusional. * Introduction Artificial intelligence is a robot that is controlled by a computer to perform task. These programs are designed to process data much faster than a human brain and make predictions. Some examples of artificial intelligence are Alexa, Siri, self-driving cars, Netflix recommendations, or even Google search algorithms. Artificial intelligence falls under two categories: narrow artificial intelligence (weak AI) and general artificial intelligence (strong AI). There are also four types of artificial intelligence: reactive machines, limited memory, theory of mind and self-awareness. Narrow artificial intelligence performs under constraints and is focused on performing a single task. There are two different types of narrow artificial intelligence, Reactive and Limited Memory. Reactive artificial intelligence doesn’t rely on large amounts of data. In fact, it doesn’t require any data storage at all as it is simply reacting to situations without the use of previous knowledge. Limited Memory artificial intelligence is much more advanced and more common and can be seen in search engines, virtual assistants like Siri and Alexa, as well as self-driving cars. Narrow artificial intelligence is performed by machine learning and deep learning. Machine learning feeds the computer data and uses statistical techniques to help it get better at tasks without having to write millions of lines for coding. Deep learning is a type of machine learning and runs input through network architecture. In the network architecture, there are hidden layers that the data travels through allowing the machine to go through “deep” learning and make input for accurate results. In theory, general artificial intelligence could carry out any task that a human could. According to Goertzal (2014) in his Journal to Artificial Intelligence, he mentions some key features of general artificial intelligence. General artificial intelligence can perform a variety of tasks and achieve a variety of goals. It should be able to handle problems differently than its creators. An intelligent system should be good at generalizing the knowledge it’s taken in and transfer that knowledge from one problem to another. There are four types of artificial intelligence. The first is called reactive machines. As mentioned before, reactive machines can’t store memory and only reacts to what’s in front of it without any prior experience or knowledge of the initial action. Because of this lack of anticipation and predictive technology, these machines can be easily fooled. The other kind of AI is known as limited memory AI. We can generate predictions based on past facts stored in limited memory. In contrast to reactive machines, it looks into the past to predict the future. Compared to a simple reactive machine, this system is much more versatile and intricate. Because it is able to incorporate previous patterns, react to the world around it, and predict future behaviors, limited memory artificial intelligence is found in self-driving cars. It allows the vehicle to observe their surrounding environment while at the same time, adapting to external factors that are constantly changing. Limited memory artificial intelligence systems are also the primary function of chatbots and virtual voice assistants. Many customer service systems, chatbots, and systems like siri are able to continually build upon their knowledge base and use this preexisting knowledge in future instances, unlike reactive machines. Limited memory artificial intelligence probably plays the biggest role in our daily lives through advertising. Anytime you are using Facebook, Instagram, google search engines, or essentially any website in general, you are taking part in limited memory artificial intelligence. Companies have been granted the permission to use your search history as well as online transactions to predict what you will buy in the future. If a customer were to buy a product on amazon, they will likely see an advertisement for a similar product in the near future because these machines are able to use prior knowledge as well as predict future behaviors to present you with a product you are likely to buy. The third type of artificial intelligence is theory of mind. This is something that hasn’t been created yet because it must understand human emotion. They would use that information on emotion and make decisions on their own. This would create a two-way relationship between artificial intelligence and humans. It even goes as deep as detecting one’s tone of voice. We’re likely to see this form of artificial intelligence in future telemarketing systems and other types of phone calls that are currently perceived as “spam.” In theory, instead of a person picking up the phone to an automated robot and likely hanging up in irritation, the artificial intelligence system would be able to detect frustration and react to it in an attempt to keep one on the phone to continue the transaction or whatever is taking place (Tegmark, 2015). A major difference in the theory of mind artificial intelligence and limited memory is that limited memory is simply using previous experiences to respond to situations, while theory of mind is actually incorporating common sense, social dynamics, and moral norms in its decision making. The last type of artificial intelligence is self-awareness. The goal is for artificial intelligence to become self-aware. It would understand what humans need without humans saying it but how they communicate it with their emotions, hand gestures, or eye contact. This requires much work into how the human consciousness works and how to program it into the machine so it can understand the human. This is an extremely difficult task as experts are still debating what consciousness actually means (Wise, & Wise, 2019). Despite the seemingly constant searching for how self-awareness and consciousness collaborate, self-aware artificial intelligence machines have the possibility to essentially incorporate pieces of the other artificial intelligences. It will be responding to actions, using previous knowledge of topics, events, and transactions, use human emotions in its responses, as well as being aware of itself along with these other functions. There are four different types of expert systems: MYCIN, DENDRAL, PXDES, and CaDet. The first expert system, MYCIN, can identify bacteria that could cause acute infections. PXDES predicts lung cancer and what degree it is in. CaDet identifies types of cancers and what stage it is in. Some advantages of an expert system are accuracy in problem solving and decision making, reliability, cost effectiveness and time reasonableness (Johnson, 2019). An expert system is more capable of solving problems at a lower time than a human, as well as being more accurate and making fewer mistakes. It is also more reliable than a human because it’s not perishable, doesn’t take time off for vacations or sick leaves. The following table is from Guru99 that explains the major differences between a human expert and an expert system: Human Expert Artificial Expertise Perishable Permanent Difficult to Transfer Transferable Difficult to Document Easy to Document Unpredictable Consistent Expensive Cost effective System * Background Mathematicians, logisticians, and programmers have been interested in artificial intelligence for a long time. They were interested in building something that could accurately solve problems at a faster rate than a human being. This eventually led to calculation and computer devices. From the 1900’s to the 1950’s there were many events that took place in artificial intelligence. In 1929, Makoto Nishimura built the first robot in Japan. This robot took knowledge from people and nature and could turn its head as well as change facial expressions. An inventor and physicist, John Vincent Atanasoff, created the Antanasoff Berry Computer (ABC). This computer weighed over 700 pounds and could solve up to 29 linear equations. This early version of the computer, while not considered a “general-purpose computer,” implemented 3 of the more primary systems that are used in modern day computers (Bunker, 2021). It would use binary digits, what are commonly known as ...