http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 ©INOSR PUBLICATIONS International Network Organization for Scientific Research ISSN: 2705-165X

The Future of Computing and (AI)

Xiuying Li Jing, Dave Albert and Chao Lei

Computer Science and Technology Jiangsu University, China ABSTRACT This study reviews the future of computing computing, and the Internet of Things and Artificial Intelligence. In computer (IoT). These data make great dataset for science, artificial intelligence (AI), training AI systems. In a few years time, AI sometimes called machine intelligence, is will touch nearly all the industries on this intelligence demonstrated by machines, in planet and there are plenty of ways AI is contrast to the natural intelligence and can transform certain industries. The displayed by humans. Leading AI textbooks automation by AI can bring about the define the field as the study of "intelligent unemployment crisis among humans since agents": any device that perceives its AI can outperform us in more and more environment and takes actions that fields. Another concern is the potential maximize its chance of successfully abuse of AI by hackers, rebel states, and achieving its goals. mathematicians in attackers. The advent of computers antiquity. The study of mathematical logic propelled society and spurred into action a led directly to Alan Turing's theory of lot of businesses and industries that would computation, which suggested that a otherwise have been impossible. machine, by shuffling symbols as simple Computers and computer technology as "0" and "1", could simulate any changed the way humans worked and conceivable act of mathematical functioned. The very first computer was a deduction. This insight, that digital behemoth. Today, everything that you computers can simulate any process of need is available on the computer through formal reasoning, is known as the Church– the internet and even the most complex Turing thesis. Today, data is streaming in computations can be carried out by these real time, fueled by the booming growth of modern wonders. digital devices, social media, cloud Keywords: Computing, Artificial, Intelligence, (AI). INTRODUCTION In computer science, artificial intelligence communicate information. It includes (AI), sometimes called machine development of both hardware and intelligence, is intelligence demonstrated software. Computing is a critical, integral by machines, in contrast to the natural component of modern industrial intelligence displayed by humans. technology. Major computing disciplines Leading AI textbooks define the field as include computer engineering, software the study of "intelligent agents": any engineering, computer science, device that perceives its environment and information systems, and information takes actions that maximize its chance of technology. successfully achieving its goals [1]. Computer vision, or the ability of Colloquially, the term "artificial artificially intelligent systems to “see” intelligence" is often used to describe like humans, has been a subject of machines (or computers) that mimic increasing interest and rigorous research "cognitive" functions that humans for decades now. As a way of emulating associate with the human mind, such as the human visual system, the research in "learning" and "problem solving" [2]. the field of computer vision purports to Computing is any activity that uses develop machines that can automate tasks computers to manage, process, and that require visual cognition. However,

1 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 the process of deciphering images, due to computing somewhat differently from a the significantly greater amount of multi- software engineer. Regardless of the dimensional data that needs analysis, is context, doing computing well can be much more complex than understanding complicated and difficult. Because society other forms of binary information. This needs people to do computing well, we makes developing AI systems that can must think of computing not only as a recognize visual data more complicated. profession but also as a discipline. But, the use of and artificial As machines become increasingly neural networks is making computer capable, tasks considered to require vision more capable of replicating human "intelligence" are often removed from the vision. In fact, computer vision is definition of AI, a phenomenon known as becoming more adept at identifying the AI effect.[3] A quip in Tesler's patterns from images than the human Theorem says "AI is whatever hasn't been visual cognitive system [3] [4]. For done yet."[7] For instance, optical instance, in the field of healthcare, character recognition is frequently computer vision-based technology has excluded from things considered to be AI, said to have exceeded the pattern having become a routine technology [8]. recognition capabilities of human Modern machine capabilities generally physicians. Researchers have tested an AI classified as AI include successfully that can detect neurological illnesses by understanding human speech [9]. reading CT scan images faster than competing at the highest level in strategic radiologists. game systems (such as chess and Go) [10]. With similarly astounding feats by AI with autonomously operating cars, intelligent computer vision technology becoming routing in content delivery networks, and increasingly common in different military simulations. industries, the future of computer vision Artificial intelligence was founded as an appears to be full of promise and academic discipline in 1956, and in the unimaginable outcomes. Read on to know years since has experienced several waves the state of computer vision technology of optimism,[11] [10] followed by today and where it is heading in the disappointment and the loss of funding future. [5] defined "computing" as follows: (known as an "AI winter"), [12] [13] in a general way, we can define followed by new approaches, success and computing to mean any goal-oriented renewed funding. [14] [15] For most of its activity requiring, benefiting from, or history, AI research has been divided into creating computers. Thus, computing subfields that often fail to communicate includes designing and building hardware with each other.[16] These sub-fields are and software systems for a wide range of based on technical considerations, such purposes; processing, structuring, and as particular goals (e.g. "robotics" or managing various kinds of information; ""),[17] the use of doing scientific studies using computers; particular tools ("logic" or artificial neural making computer systems behave networks), or deep philosophical intelligently; creating and using differences. Subfields have also been communications and entertainment based on social factors (particular media; finding and gathering information institutions or the work of particular relevant to any particular purpose, and so researchers) [18]. on. The list is virtually endless, and the The traditional problems (or goals) of AI possibilities are vast. research include reasoning, knowledge However, [6] also recognizes that the representation, planning, learning, natural meaning of "computing" depends on the language processing, perception and the context: ability to move and manipulate Computing also has other meanings that objects.[19] General intelligence is among are more specific, based on the context in the field's long-term goals [20]. which the term is used. For example, an Approaches include statistical methods, information systems specialist will view computational intelligence, and

2 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 traditional symbolic AI. Many tools are explored by myth, fiction and philosophy used in AI, including versions of search since antiquity.[22] Some people also and mathematical optimization, artificial consider AI to be a danger to humanity if neural networks, and methods based on it progresses unabated. Others believe statistics, probability and economics. The that AI, unlike previous technological AI field draws upon computer science, revolutions, will create a risk of mass information engineering, mathematics, unemployment [23]. psychology, linguistics, philosophy, and In the twenty-first century, AI techniques many other fields. have experienced a resurgence following The field was founded on the assumption concurrent advances in computer power, that human intelligence "can be so large amounts of data, and theoretical precisely described that a machine can be understanding; and AI techniques have made to simulate it" [21]. This raises become an essential part of the philosophical arguments about the nature technology industry, helping to solve of the mind and the ethics of creating many challenging problems in computer artificial beings endowed with human-like science, software engineering and intelligence. These issues have been operations research. THEORETICAL REVIEW The study of mechanical or "formal" knowing, blazing fast and contextually reasoning began with philosophers and aware. mathematicians in antiquity. The study of The odd of that computer being an mathematical logic led directly to Alan aluminum clamshell is nil. Understanding Turing's theory of computation, which this transformation is key. It explains why suggested that a machine, by shuffling Alphabet, Microsoft and Amazon are symbols as simple as "0" and "1", could betting their future on powerful cloud simulate any conceivable act of computing networks and artificial mathematical deduction. This insight, intelligence, the branches of computer that digital computers can simulate any science that teach computers to learn like process of formal reasoning, is known as humans by recognizing and the Church–Turing thesis.[24] Along with understanding patterns. concurrent discoveries in neurobiology, These companies also have key scale information theory and cybernetics, this advantages that will ensure future led researchers to consider the possibility profitability. Building large cloud and AI of building an electronic brain. Turing environments is prohibitively expensive proposed changing the question from for smaller enterprises. Training AI sucks whether a machine was intelligent, to up a lot of computing power. In March, "whether or not it is possible for Recode reported the triumvirate spent a machinery to show intelligent record $20 billion in 2017 total to equip behaviour".[25] The first work that is now and build new data centers. The outlay generally recognized as AI was was more than the previous three years McCullouch and Pitts' 1943 formal design combined. Let that sink in for a moment. for Turing-complete "artificial neurons" At the same time, they are leading a gold [26]. rush of new AI investment. In December The future of computing is software, not 2017, The Economist reported the largest hardware. We are living in an age in which technology companies had acquired 110 powerful cloud computing networks make AI companies since 2010. The value of supercomputing ubiquitous and relatively those investments surged beyond $22 inexpensive. Within two years, 5G billion. wireless networks will kill latency. Add At Alphabet, machine learning is in its optical, audio, GPS and other sensors corporate DNA. Machine learning is trial- commoditized by smartphone mass and-error on a massive scale. Legions of production, and you have the ingredients computers crunch through enormous sets for a computing device that is all- of data, learning through progression. Since 1994, the company has depended

3 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 on it for Google Search and the It’s important for investors to look development of its digital advertising forward. The future of computing is not model. Today, it is the engine that powers complicated, and it is going to be a big every service at Alphabet In 2016, Jeff business. Thankfully, it is all being Bezos, chief executive officer at Amazon, carefully laid out right now. The field of declared AI is its fourth pillar, behind AI research was born at a workshop at ecommerce, Amazon Prime, its Dartmouth College in 1956,[29] where the membership service, and Amazon Web term "Artificial Intelligence" was coined Services, its cloud division. And in 2017, by John McCarthy to distinguish the field Satya Nadella, chief executive officer at from cybernetics and escape the influence Microsoft, changed its corporate mission of the cyberneticist Norbert Wiener.[30] from mobile and cloud first, to AI first. Attendees Allen Newell (CMU), Herbert The new strategy involved building AI Simon (CMU), John McCarthy (MIT), Marvin into every software application and Minsky (MIT) and Arthur Samuel (IBM) service to ensure the user experience became the founders and leaders of AI spanned every device [27]. research.[31] They and their students All three companies have made voice a produced programs that the press cornerstone of their AI efforts. Developers described as "astonishing":[32] computers can build applications for Google were learning checkers strategies (c. Assistant, Alexa and Microsoft Cortana 1954)[33] (and by 1959 were reportedly that run everywhere. It is not hard to see playing better than the average the future of computing. It begins with a human),[34] solving word problems in personal digital assistant that lives in the algebra, proving logical theorems (Logic cloud. That assistant is unique to its user Theorist, first run c. 1956) and speaking and will have access to all personal data. English.[35] By the middle of the 1960s, It will also be in a constant state of research in the U.S. was heavily funded by refinement. It will learn. the Department of Defense[36] and It’s the future promised in the 2013 film, laboratories had been established around “Her.” The Spike Jonze dark romantic the world.[37] AI's founders were comedy is the story of a man and his love optimistic about the future: Herbert affair with his AI digital assistant. It is a Simon predicted, "machines will be strange concept. But, it is not too far from capable, within twenty years, of doing any the realm of possibility given how quickly work a man can do". AI is progressing. agreed, writing, "within a generation the [28], the noted futurist and chief engineer problem of creating 'artificial intelligence' at Google, was a technical adviser to the will substantially be solved". [8] film. Today, Kurzweil is tasked with They failed to recognize the difficulty of bringing AI to Gmail, the popular email some of the remaining tasks. Progress client. AI networks quickly scan every slowed and in 1974, in response to the email, looking for keywords and context, criticism of Sir James Lighthill [18] and before offering an automated response. It ongoing pressure from the US Congress to is all still rudimentary, but you can see fund more productive projects, both the the direction. U.S. and British governments cut off The company is using billions of emails exploratory research in AI. The next few as data to refine AI algorithms that years would later be called an "AI understand written context. It is a small winter",[10] a period when obtaining first step toward digital assistants capable funding for AI projects was difficult. of conversation. Amazon uses AI to tailor The Future Is Now: Ai's Impact Is recommendations at its giant online Everywhere marketplace and to choreograph its There’s virtually no major industry 80,000 warehouse robots. And Microsoft modern AI more specifically, “narrow AI,” is working with partners to build always which performs objective functions using on thin and light computers enhanced by data trained models and often falls into Cortana. the categories of deep learning or

4 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 machine learning hasn’t already affected. Nowadays private firms are becoming the That’s especially true in the past few protagonists rather than operating as a years, as data collection and analysis has contractor in the space industry [29]. One ramped up considerably thanks to robust of the notable examples is of SpaceX, in IoT connectivity, the proliferation of 2018 they launched 21 rockets in the connected devices and ever-speedier space [10]. Blue Origin plans to start computer processing. Some sectors are at space tourism by 2021. NASA is using AI the start of their AI journey, others are for trajectory and payload optimization to veteran travelers. Both have a long way to increase the efficiency of the next rover go. Regardless, the impact artificial mission to Mars [21]. AI will extend the intelligence is having on our present day boundaries of human compatibilities and lives is hard to ignore. will help scientist to reach Europa, Applications of Artificial Intelligence in Jupiter’s moon where scientist believe future technologies there could be a subsurface ocean. NASA Today, data is streaming in real time, is currently planning to launch James fueled by the booming growth of digital Webb Space Telescope in the orbit around devices, social media, cloud computing, 1.5 million kilometers from the earth in and the Internet of Things (IoT). These 2020, part of the mission will be data make great dataset for training AI overlooked by the AI. AI is still gathering systems. In a few years time, AI will touch momentum in the space industry [1]. The nearly all the industries on this planet coming years’ mission will be and there are plenty of ways AI is and can turbocharged with AI as we voyage in transform certain industries. comets, planets, moons and explore the Health Care possibility of mining the comets [2]. AI system can help medical physicians Environmental Protection deliver faster and accurate treatment by With the impending unprecedented stress analyzing a large amount of clinical data caused by global warming, natural like demographics, medical notes, disasters, and other human activities, AI recordings from medical devices and can help us take concrete steps to better laboratory images [8]. For instance, AI can protect our environment. For instance, by help in detecting epithelial ovarian cancer using Youtube videos to uniquely identify in stage 1A when it has a 94% cure rate and track the movement of animals, AI [15]. Whereas, by a normal process, it is can help scientists to identify and protect usually undetected or is detected in stage the endangered animals [33]. In future AI 3 or 4, when the symptoms start can help in greenhouse gas reduction appearing and chances of cure start through proper traffic optimization, route plummeting. Moreover, Elon Musk’s planning of autonomous public transport Neural Lace could be the next AI and ride-sharing services [4]. It will advancement in the field of healthcare. It augment agriculture with the help of is an ultra-thin mesh that can be automated data collection and take embedded in the skull creating an corrective action using robots to detect interface between the machine and the crop diseases which will eventually brain. Gradually, it would become a part increase the efficiency and reduce the use of the brain and help in treating brain of pesticides and fertilizers. AI will also disorders [26]. Researchers are hoping help in monitoring the coral reefs by that in future AI may enhance the ability processing the plethora of images of a human to provide better healthcare collected by drones and NASA’s satellites services [5] and will enhance life [24]. This information can help in expectancy of human civilization. protecting the reefs from collapse. The Space Industry intelligence and productivity gains that AI Historically machine learning algorithm will deliver can unlock robust solutions to has been used in health monitoring of the environment’s pressing challenges spacecraft, navigation, intelligent control like these. and object detection for navigation.

5 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 Artificial General Intelligence accidents and better fuel efficiency due to Highly specialized machines like Deep smoother braking and acceleration than Blue, Watson and AlphaGo are single- human drivers. In the military, by purposed AI systems whose intelligence is harnessing satellite photo interpretation very domain-specific known as Artificial capabilities, AI programs could identify Narrow Intelligence(ANI) [35]. There is potential targets and threats. The next ongoing experimental work on Artificial evolution that we might see in AI is General intelligence that would be flexible Neuromorphic Computing, which will be enough to learn without supervision and similar to the human brain model. AI can adapt to multiple and unexpected be harnessed further to help in getting rid situations like humans. In the field of of the errors that human brains are prone legal systems, it would implement to make [11]. Moreover, we might see the concepts like “fairness” and “justice”. integration of quantum hardware and After decades, emotional robots won’t be software with AI, to solve complex a thing of sci-fi fantasy. problems within seconds [29]. Others Challenges of Artificial Intelligence In other fields like cybersecurity, The automation by AI can bring about the Artificial Neural Network will play a unemployment crisis among humans significant role by consistently since AI can outperform us in more and accumulating intelligence on attacks, more fields. Another concern is the breaches, new threats, malware and given potential abuse of AI by hackers, rebel all this information AI can learn from it states, and attackers. For instance, a and detect abnormalities within an commercial drone can be turned into a organization’s network and flag it quicker targeted weapon. Moreover, Scientists than a member of cybersecurity [26]. believe after Artificial General Harnessing AI, chat-bots will be more Intelligence, next we might see is effective which will make a great Artificial Super Intelligence, which would transition from the Graphical User surpass human intelligence. Such super Interface (GUI) to the Conversational User intelligence might rule the world and see Interface (CUI) [17]. Moreover, Automated humans as a threat and a waste of vehicles will be tremendously substituted resource. by driverless cars that would reduce road CONCLUSION The advent of computers propelled tiniest of devices and commodities, might society and spurred into action a lot of soon have computers embedded within businesses and industries that would them. There is an increasing demand for otherwise have been impossible. products that can carry out the Computers and computer technology computations that they need on their own changed the way humans worked and and this has led to the need for devices functioned. The very first computer was a that have computers embedded in them. behemoth. Today, everything that you Some people perceive the final goal of need is available on the computer through computers to be that of ensuring that the internet and even the most complex computers are inextricable from society computations can be carried out by these and human life. Computers would be so modern wonders. Computers are already involved in every process that the very ubiquitous, but, they possess the ability thought of its non-existence would make to enhance their presence even more. life impossible! The aim is to ensure that Computers are on desks and countertops, computers are entangled with human life in bags and pockets, but, soon, they in such a way that computers would be might be a part of everything imaginable. indistinguishable from human life. Yes, everything conceivable, even the

6 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 REFERENCES 1. Maloof, Mark. "Artificial said "Artificial intelligence is not, Intelligence: An Introduction, p. by definition, simulation of human 37"(PDF). georgetown.edu. intelligence" (Maker 2006). 2. Schank, Roger C. (1991). "Where's 16. Ford, Martin; Colvin, Geoff the AI". AI magazine. Vol. 12 no. 4. (2015). "Will robots create more p. 38. jobs than they destroy?". The 3. ^ Jump up to:a b Optimism of Guardian. Retrieved 13 early AI: January 2017. 4. Herbert Simon quote: Simon 1965, 17. Russell & Norvig 2003, p. 939 p. 96 quoted in Crevier 1993, 18. McCorduck 2004, pp. 17–25 p. 109. 19. Berlinski, David (2000). The Advent 5. Marvin Minsky quote: Minsky of the Algorithm. Harcourt 1967, p. 2 quoted in Crevier 1993, Books. ISBN 978-0-15-601391- p. 109. 8. OCLC 46890682. 6. Boom of the 1980s: rise of expert 20. Turing, Alan (1948), "Machine systems, Fifth Generation Intelligence", in Copeland, B. Jack Project, Alvey, MCC, SCI: (ed.), The Essential Turing: The 7. McCorduck 2004, pp. 426–441 ideas that gave birth to the 8. Crevier 1993, pp. 161–162,197– computer age, Oxford: Oxford 203, 211, 240 University Press, p. 412, ISBN 978- 9. Russell & Norvig 2003, p. 24 0-19-825080-7 10. Pamela McCorduck (2004, pp. 424) 21. Russell & Norvig 2009, p. 16. writes of "the rough shattering of 22. Russell & Norvig 2003, p. 17, who AI in subfields vision, natural write "for the next 20 years the language, decision theory, genetic field would be dominated by these algorithms, robotics and these people and their students." with own sub-subfield that would 23. McCorduck 2004, pp. 129–130 hardly have anything to say to 24. Russell & Norvig 2003, p. 18. each other." 25. Schaeffer J. (2009) Didn't Samuel 11. Poole, Mackworth & Goebel 1998 Solve That Game?. In: One Jump 12. Nilsson 1998 Ahead. Springer, Boston, MA 13. Russell & Norvig 2003, pp. 2–3, 26. Samuel, A. L. (July 1959). "Some who make the analogy Studies in Machine Learning Using with aeronautical engineering. the Game of Checkers". IBM Journal 14. McCorduck 2004, pp. 100–101, of Research and who writes that there are "two Development. 3 (3): 210– major branches of artificial 229. CiteSeerX 10.1.1.368.2254. do intelligence: one aimed at i:10.1147/rd.33.0210. producing intelligent behavior 27. Russell & Norvig 2003, pp. 18–21 regardless of how it was 28. Luger & Stubblefield 2004, accomplished, and the other aimed pp. 227–331 at modeling intelligent processes 29. Nilsson 1998, chpt. 17.4 found in nature, particularly 30. Clark, Jack (8 December human ones." 2015). "Why 2015 Was a 15. Kolata 1982, a paper in Science, Breakthrough Year in Artificial which describes McCarthy's Intelligence". Bloomberg indifference to biological models. News. Archived from the original Kolata quotes McCarthy as writing: on 23 November 2016.. "This is AI, so we don't care if it's 31. Lorica, Ben (2017). "The state of AI psychologically real""Science". adoption". O'Reilly Media. August 1982.. McCarthy recently Retrieved 2 May 2018. reiterated his position at 32. Allen, Gregory the AI@50 conference where he (2017). "Understanding China's AI

7 http://www.inosr.net/inosr-applied-sciences/ Xiuying et al INOSR APPLIED SCIENCES 3(1): 1-8, 2017 Strategy". Center for a New 36. Matti, D.; Ekenel, H. K.; Thiran, J. P. American Security. (2017). Combining LiDAR space 33. Kaplan, Andreas; Haenlein, Michael clustering and convolutional (2017). "Siri, Siri, in my hand: neural networks for pedestrian Who's the fairest in the land? On detection. 2017 14th IEEE the interpretations, illustrations, International Conference on and implications of artificial Advanced Video and Signal Based intelligence". Business Surveillance (AVSS). pp. 1–6. Horizons. 62 (1): 15–25 37. Ferguson, Sarah; Luders, Brandon; 34. Lindenbaum, M., Markovitch, S., & Grande, Robert C.; How, Jonathan Rusakov, D. (2004). Selective P. (2015). Real-Time Predictive sampling for nearest neighbor Modeling and Robust Avoidance of classifiers. Machine learning, 54(2), Pedestrians with Uncertain, 125–152. Changing Intentions. Algorithmic 35. Goodfellow, Ian J., Jonathon Foundations of Robotics XI. Shlens, and Christian Szegedy. Springer Tracts in Advanced "Explaining and harnessing Robotics. 107. Springer, Cham. adversarial examples." arXiv pp. 161–177. preprint arXiv:1412.6572 (2014).

8