Computer Science Capstone Project Examples

COMPUTER SCIENCE CAPSTONE EXAMPLE ARTIFICIAL INTELLIGENCE TODAY               INTRODUCTION               Efforts to build artificial intelligence have experienced a striking rise in recent yearswith numerous investments in the application and development of intelligent systems inareas such as communications, commerce, health services, internet search, referralsystems, etc. While numerous scientific fantasy works entertain themselves by creatingsome form of consciousness artificial intelligence, it seems that the first signs of intelligentmachines are already present in computers that learn, recognize objects, and use thelanguage of communicating to people better. The paper deals with artificial intelligencetoday, i.e. the ability of computers to behave in an intelligent way. The main aim of the paperis to describe the technology of artificial intelligence that is present today and to giveexamples in which it is used. The reason for choosing a theme is the fact that in the last fewyears the development of machine learning and the application of intelligent systems havebeen growing steadily, but the wider public knows very little about it. Today we can safelysay that artificial intelligence will mark the next decade, launch a new technologicalrevolution and change our lives in the way that in this the moment we cannot fully imagine.              Artificial Intelligence (UI) is the ability of a digital computer or a computer-controlled robot to perform tasks commonly associated with intelligent creatures. The fieldof artificial intelligence, shortened UI, seeks to understand intelligent entities. So one of thereasons for studying this area is to better understand ourselves. However, as opposed tophilosophy or psychology, which also studies intelligence, the field of artificial intelligenceseeks the construction of artificial entities as well as their understanding. Artificialintelligence asks puzzles. How can a slow or small brain, biologically or electronically, beable to perceive, understand and anticipate the world, and manipulate the world much largerand more complex than it is? How do we build something with such properties? Thesequestions are tough, but unlike traveling at speeds greater than the speed of light orantigravity, the researcher in the field of artificial intelligence has strong evidence that thetask can be accomplished. All we have to do is look into the mirror to see an example of anintelligent system. There are two main goals sought by most artificial intelligenceresearches. The first, most important goal is to build intelligent machines. Another goal is tounderstand the nature of intelligence or what psychologists call the g-factor in intelligencetests (an attempt to measure general intelligence spreading through different human  1

COMPUTER SCIENCE CAPSTONE EXAMPLE ARTIFICIAL INTELLIGENCE TODAYdomains). Both goals have in their essence the need to define the notion of intelligence.               General Artificial Intelligence GUI is a term that describes research aimed at creatingmachines capable of generally intelligent operations. It is crucial that general artificialintelligence must be able to master different areas and learn new areas that they have notfaced before. It is not necessary to have equal capabilities in all domains. When in thatcontext we say human mind or GUI have a general purpose, we do not think that any problemcan be solved in every domain, but that there is a potential for solving any problem in anydomain if a proper experience is obtained. As long as it does not prove that GUI is notpossible, it remains a legitimate topic of research. And considering the complexity of theproblem, there is no reason to expect the GUI to achieve the goal in a short period, so allpopular theoretical controversies will probably even exist after the GUI is realized asplanned. The fact that research has little consensus needs to make us more cautious whenwe think of a new idea as completely wrong. As history has already happened more thanonce, the real penetration could come from something that opposes the intuition.              In recent years, technology companies and academic researchers are trying tocreate a so-called neuromorphic computer architecture - that is, chips that emulate theability of the human brain to be both analytical and intuitive to enable the creation of contextand meaning from a large amount of data. Scientists who conduct such research are calledneuromorphic engineers. Instead of thinking about the brain as a computer, they are tryingto create computers that resemble the brain. In this way, humanity will gain not only a betterunderstanding of brain work, but better, more intelligent accounts. While the human brainhas 100 trillion of synapses and consumes only 20W, today's superfine in the effort tosimulate brain work consumes the force of the order of magnitude MW. Leading efforts toachieve a system whose properties are similar to the brain have reached a new milestone byproducing a transistor chip containing more than 4000 neurosinaptic nuclei. Each coreconsists of computational components corresponding to their biological duality - the corememory function similar to are synapses among neurons, processors represent coreneurons, and communication is accomplished by conductors similar to neural axons. Thegoal of non-neurological scientists is to build a computer that has some or all of the featuresthat a brain has, and today's computers do not have that low energy consumption; faulttolerance (failure of a transistor creates serious problems in the microprocessor, but thebrain permanently loses some neurons, which does not cause problems in the functioning 2

COMPUTER SCIENCE CAPSTONE EXAMPLE ARTIFICIAL INTELLIGENCE TODAYof the nervous system) and the need for programming (brains learn and changespontaneously through their interaction with the world instead of following the default pathsand branches of the predetermined algorithm). To achieve this goal, scientists should knowbrain work, but it is still a big unknown. There is no way they could, for now, be able to studythe brain at a fundamental level. On the other hand, suitable computer simulations couldanswer some questions about basic brain functions and vice versa. For right, achieving agood brain imaging would have turned into a computer literacy that could ultimately servebetter brain function, artificial intelligence development, and possibly conscientiouscomputer systems. It might happen that the models are first and then help brain mapping.Neuromorphic engineering could, in other words, reveal the underlying principles of thinkingbefore neuroscientist do. The large gap in brain incomprehension is in the middle scale, or inthe middle degree of brain anatomy. Science is known for the work of individual neurons. Itis also relatively well known how individual brain neurons and ganglia (neuronal clustersconstructing a peripheral nervous system) work, where speech or vision centers are locatedin the brain. However, it is unclear how neurons are organized in the ganglia and the braintrays, and that is precisely the level of organization in which the thinking is realized - and it isassumed that consciousness is inhabited.What is increasingly mentioned is IoT (Internet of Things). It is accomplished when we takeobjects from everyday life and add to them the ability to collect information (about ourcustom habits), communicate with each other and with us. That term does not includepersonal computers, smartphones, and tablets, so it is only about everyday things like homeappliances. The predictions indicate that the growth of the IoT will exceed the otherconnecting devices.  Connecting things, the flood of information gathered from the sensor,which will be cheaper and smaller, will require intelligent systems that will manage thedevices inexperienced users in some future smart cities. Personalization is a solution to theworld's overwhelming information. Personalization is carried out by applying intelligentsystems to data collected by users. Data is collected over the internet on social networks orthrough what the user enters in the search engine or from a variety of sensors, trackingsystems or credit cards. Search results tailored first to the user's personal interests, tailoredto the user's habits and preferences, customized user news, filtering \"relevant\" topics, andcustomizing advertiser bids for consumer habits. All this is based on the ability of smartsystems to process and interpret the collected data. The ad customization carried out by the  3

COMPUTER SCIENCE CAPSTONE EXAMPLE ARTIFICIAL INTELLIGENCE TODAYadvertiser's bid is based on conclusion smart software about our intentions throughsampling patterns from sites we buy or based on our social media releases.REFERENCESBell, J. (2014). Machine Learning: Hands-On for Developers and Technical Professionals, NewYork: John Wiley & Sons Inc.Bostrom, N. (2014). Superintelligence, Oxford: Oxford University Press.Kurzweil, R. (2012). How to Create a Mind: The Secret of Human Thought Revealed, New York:Viking Penguin.Pesaran, A., Gonder, J., & Keyser, M. (2009). Ultracapacitor Applications and Evaluation forHybrid Electric Vehicles, In: 7th Annual Advanced Capacitor World Summit Conference,National Renewable Energy Laboratory (NREL): Hotel Torrey Pines La Jolla, CA; 2009.Tie, S. F., & Tan, C. W. (2013). A Review of Energy Sources and Energy Management Systemin Electric Vehicles. Renewable and Sustainable Energy Reviews, 20, 82-102.Emadi, A., Young, J. L., & Rajashekara, K. (2008). Power Electronics and Motor Drives inElectric, Hybrid Electric, and Plug-In Hybrid Electric Vehicles. Industrial Electronics, IEEETransactions on 2008, 55(6): 2237–45. 4