Applications and Pros and Cons of Artificial Intelligence

Abstract

Artificial intelligence (AI), deep learning, system getting to know and neural networks constitute incredibly exciting and powerful machine studying-based strategies used to resolve many real-international problems. For a primer on gadget learning. While human-like deductive reasoning, inference, and decision-making by means of a laptop is nonetheless a long time away, there have been top notch gains in the application of AI techniques and related algorithms.

Introduction

Artificial intelligence is a department of pc technology that targets to create smart machines. It has emerged as an essential a part of the era industry. Research related to artificial intelligence is quite technical and specialized. The core troubles of synthetic intelligence include programming computers for sure traits such as: Knowledge, Reasoning, Problem-solving, Perception, Learning, Planning, and Ability to manipulate and pass objects. Knowledge engineering is a core a part of AI research.

Classification determines the category an object belongs to and regression offers with obtaining a set of numerical enter or output examples, thereby coming across functions allowing the generation of suitable outputs from respective inputs. Mathematical evaluation of device mastering algorithms and their performance is a well-described branch of theoretical computer technological know-how often referred to as computational gaining knowledge of theory. Machine notion offers with the capability to apply sensory inputs to deduce the different components of the world, while pc imaginative and prescient is the strength to analyze visual inputs with a few sub- problems consisting of facial, object and gesture recognition. Robotics is also a chief field related to AI. Robots require intelligence to deal with tasks including object manipulation and navigation, at the side of sub-issues of localization, motion making plans and mapping.

Deep gaining knowledge of whilst flashy is certainly just a term to describe certain styles of neural networks and associated algorithms that consume frequently very raw input data. They system this information through many layers of nonlinear alterations of the input statistics so one can calculate a goal output.

Unsupervised function extraction is also an area where deep mastering excels. Feature extraction is while an algorithm is capable of automatically derive or assemble meaningful features of the information to be used for further gaining knowledge of, generalization, and understanding. The burden is traditionally on the facts scientist or programmer to perform the function extraction system in maximum other device learning approaches, along with characteristic choice and engineering.

Feature extraction usually involves some quantity dimensionality discount as well, which is reducing the amount of input functions and data required to generate meaningful results. This has many benefits, which consist of simplification, computational and memory energy reduction, and so on. Programmers would teach a neural network to locate an object or phoneme by means of blitzing the community with digitized variations of pics containing the ones objects or sound waves containing the ones phonemes. If the network didn’t accurately recognize a particular pattern, a set of rules would modify the weights. The eventual aim of this education changed into to get the network to consistently apprehend the styles in speech or sets of photographs that we humans recognize as, say, the phoneme “d” or the photograph of a dog. This is a whole lot the same manner a infant learns what a canine is by way of noticing the info of head shape, behavior, and so forth in furry, barking animals that other human beings call dogs.

Machine gaining knowledge of came directly from minds of the early AI crowd, and the algorithmic methods over the years included choice tree learning, inductive logic programming. Clustering, reinforcement learning, and Bayesian networks among others. As we know, none performed the closing purpose of General AI, and even Narrow AI was on the whole out of attain with early system gaining knowledge of techniques.

As it turned out, one of the very exceptional application regions for gadget studying for many years became laptop vision, although it still required a exceptional deal of hand-coding to get the process done. People would pass in and write hand-coded classifiers like facet detection filters so the program should discover in which an object commenced and stopped; shape detection to decide if it had eight sides; a classifier to understand the letters “S-T-O P.” From all the ones hand-coded classifiers they would increase algorithms to make sense of the photo and “learn” to determine whether it become a prevent sign.

Good, however no longer mind-bendingly great. Especially on a foggy day when the sign isn’t perfectly visible, or a tree obscures a part of it. There’s a reason pc imaginative and prescient and picture detection didn’t come close to rivaling humans until very recently, it became too brittle and too at risk of error. Time, and the right gaining knowledge of algorithms made all the difference.

Concept of Automation

The use of automation began to growth inside the ultimate decade with an intention to lessen manpower and time. Automation has brought a device of computer and machines and replaced a machine that become built by using combining guy and machine. Highly intense and repetitive tasks have turn out to be green and the product best has also elevated with the use.

Data: Automation can also or won’t be primarily based on synthetic intelligence. The entire practice of automation has advanced into its current form between the primary and third industrial revolution. It includes production the use of automatic testing, mechanical labor, control systems, computer systems and operating equipment’s. All the types of automation which has manifested all around us are bound the use of express programming and rules. To make certain that the equal thing will become an AI, all this is had to be performed is to electricity it up the use of data. Huge quantities of data, like the usage of neural networks, graphs and deep system learning should be put in the software. Your coding stages will really determine simply how much you may be able to make your gadget stimulate like a human. But most likely, you’ll be teaching the device all which you already understand. In case of automated, you’ll be able to easily recognize the output the use of sensor readings. But in case of AI there is constantly a little bit of uncertainty, much like it is there with the human brain.

Purpose: Automation can execute repetitive tasks. This frees up treasured time for human beings to take up greater essential challenge which require rational judgment and thought. This makes everything greater green and fee effective. Artificial Intelligence is designed to not just are looking for patterns but also to analyze from revel in so that they can self-select the proper responses in keeping with situations.

Pros and Cons of Artificial Intelligence

AI offers reliability, cost-effectiveness, solve complicated problems, and make decisions; in addition, AI restrict data from getting lost. AI is applied nowadays in most fields whether business or engineering. One of the great tools in AI is called “reinforcement learning” which is based on testing success and failure in real life to increase the reliability of applications. Unfortunately, AI is limited with its capability and functionality. Although Artificial Intelligence made our lives much easier and saved us more time than ever, scientists are predicting that by the huge dependency on AI humanity could extinct. Scientists argue that by having a AI machines, people will be jobless and that will conclude in losing the sense of living. Since machines are learning and doing thigs more efficiently and effectively in a timely manner, this could be the reason of our extinction.

AI Algorithms and Models

AI is mainly based totally on algorithms and models as a way that is designed based totally on clinical findings such as math, statists, and biology AI works based on numerous fashions together with: Ant Colony Algorithm, Immune Algorithm, Fuzzy Algorithm, DecisionTree, Genetic Algorithm, Particle Swarm Algorithm, Neural Network, Deep Learning and in this report, I will discuss some of the maximum recognized models which are: Support Vector Machine, and the Artificial Neural Network,

· Support Vector Machine (SVM) where it is used to build a classification model by finding an optimal hyperplane based on a set of training examples as shown in (figure A-1). It is also have been used for pattern classification and trend prediction lots of applications for instance: power transformer fault diagnosis, disease diagnosis and treatment optimization. (Li& Jiang, (n.d.)).

· Artificial Neural Network (ANN) is a representative model of understanding thoughts and behaviors in terms of physical connection between neurons. ANN has been used to solve variety of problems through enabling the machine to build mathematical models to be able to imitate natural activities from brains perspective as shown in. By using this algorithm, the machine will be able to identify the solution of any problem just like human’s brain.

Some Applications on Artificial Intelligence:

AI can be designed using masses of algorithms. These algorithms help the gadget to decide the predicted response which will basically inform the computer what to anticipate and work accordingly. Here are some of the finest AI applications that we are probably the use of in each day life without knowing:

• Voice recognition
• Virtual agents
• Machine learning platform
• Ai optimized hardware
• Decision management
• Deep learning platform
• Biomatters
• Robotic process automation
• Text analytics and NLP
• Adaptive Manufacturing:
• Machines that are “able to learn a multitude of tasks from demonstrations, just like their human counterparts can.” (“Yoa”,2017))

AI Design Models

AI software are a lot around us and in this paper, I will discuss some of the maximum commonplace utility of AI that we continually use nowadays that is Virtual Assistants which includes Siri, Cortana…Etc. Over the past few years clever assistants are becoming a very commonplace technology in most of the clever gadgets and maximum importantly, that those assistants are becoming smarter than ever. In addition to the brilliant help they provide us with, is that all of those apps have particular features. Artificial Intelligence works according to the subsequent phases: getting the information, clean/manipulating/ prepare the data, teach model, test records, and improve the facts as cited in. Before having access to the information, a business must affirm the satisfactory of the statistics to ensure that it meets the requirement.

Siri Virtual Assistant:

Siri is the well-known digital assistant which makes use of voice recognition and typed commands so as to carry out a certain project inside a device. Siri is taken into consideration one in every of AI maximum used applications. The application genuinely takes the input from the consumer such as (e.g. Call dad) and attempt to locate the most associated keywords used on this command. Siri tries to get rid of inconsistent result through the use of the language pattern recognizer and from there to lively ontology via searching via the contacts, then it tries to relate the touch named “Dad” and carry out the challenge which is in this situation is “Calling” and in the end the output of this action will be “calling dad” and to recollect all of the possible conditions.

In another scenario the architecture of the virtual assistant is shown in as we can see the flow of the system starts by taking the input from the user, after that the system decides the conversation strategy module to be used which is a respond from the dialog management module, meanwhile a classification module response to an NLP module. Finally, using the conversation history database is used to analyze the knowledge base construction module which will response back to the domain knowledge based as explained.

Conclusion

AI is an extremely effective and exciting field. It’s most effective going to grow to be more essential and ubiquitous moving forward, and will certainly maintain to have very good-sized influences on modern-day society. Artificial neural networks (ANNs) and the more complicated deep learning technique are a number of the maximum successful AI equipment for fixing very complex problems, and could maintain to be evolved and leveraged within the future. While a terminator-like state of affairs is unlikely any time soon, the progression of artificial intelligence strategies and programs will certainly be very thrilling to watch!

References

1. Habeeb, A. (2017). Introduction to Artificial Intelligence. [online] ResearchGate. Available at: https://www.researchgate.net/publication/325581483_Introduction_to_Artificial_Intelligence [Accessed 28 Dec. 2019].
2. The American University of the Middle East (2017). (PDF) Artificial Intelligence. [online] ResearchGate. Available at: https://www.researchgate.net/publication/323498156_Artificial_Intelligence [Accessed 28 Dec. 2019].
3. ResearchGate. (2019). Simanta Shekhar Sarmah. [online] Available at: https://www.researchgate.net/profile/Simanta_Sarmah [Accessed 28 Dec. 2019].
4. ResearchGate. (2019). (PDF) Primer on artificial intelligence and robotics. [online] Available at: https://www.researchgate.net/publication/333469880_Primer_on_artificial_intelligence_and_robotics [Accessed 28 Dec. 2019].
5. The Impact of Artificial Intelligence on Global Trends. (n.d.). [online] Available at: https://www.researchgate.net/profile/Alice_Pavaloiu/publication/325398084_The_Impact_of_Artificial_Intelligence_on_Global_Trends/links/5b0bd55ba6fdcc8c2534cfc1/The-Impact-of-Artificial-Intelligence-on-Global-Trends.pdf?origin=publication_detail [Accessed 28 Dec. 2019].
6. University College London (2017). (PDF) Artificial Intelligence and Robotics. [online] ResearchGate. Available at: https://www.researchgate.net/publication/318858957_Artificial_Intelligence_and_Robotics [Accessed 28 Dec. 2019]
7. Li, X. and Jiang, H. (n.d.). Artificial Intelligence Technology and Engineering Applications. [online] Available at: https://pdfs.semanticscholar.org/1177/177d1531f9d6a119f589c23e9ab2e30294db.pdf.
8. Purcher, J. (2012). Apple introduces us to Siri, the Killer Patent. [online] Patently Apple. Available at: https://www.patentlyapple.com/patently-apple/2012/01/apple-introduces-us-to-siri-the-killer-patent.html [Accessed 4 Jan. 2020]
9. Knight, W. (2017). 5 Big Predictions for Artificial Intelligence in 2017. [online] MIT Technology Review. Available at: https://www.technologyreview.com/s/603216/5-big-predictions-for-artificial-intelligence-in-2017/ [Accessed 4 Jan. 2020
10. Jahnavi Mahanta (2017). Introduction to Neural Networks, Advantages and Applications. [online] Medium. Available at: https://towardsdatascience.com/introduction-to-neural-networks-advantages-and-applications-96851bd1a207 [Accessed 4 Jan. 2020]
11. London, in (2019). O’Reilly Artificial Intelligence Conference in London 2019. [online] Oreilly.com. Available at: https://conferences.oreilly.com/artificial-intelligence/ai-eu-2019 [Accessed 4 Jan. 2020].
12. Artificial Intelligence Applications to Critical Transportation Issues. (2012). [online] Available at: http://onlinepubs.trb.org/onlinepubs/circulars/ec168.pdf.
13. The History of Artificial Intelligence. (2006). [online] Available at: https://courses.cs.washington.edu/courses/csep590/06au/projects/history-ai.pdf.
14. Mariya Yao (2017). Factories Of The Future Need AI To Survive And Compete. Forbes. [online] 19 Aug. Available at: https://www.forbes.com/sites/mariyayao/2017/08/08/industrial-ai-factories-of-future/#64694ac3128e [Accessed 4 Jan. 2020].
15. The Official NVIDIA Blog. (2016). The Difference Between AI, Machine Learning, and Deep Learning? | NVIDIA Blog. [online] Available at: https://blogs.nvidia.com/blog/2016/07/29/whats-difference-artificial-intelligence-machine-learning-deep-learning-ai/ [Accessed 4 Jan. 2020].
16. Google Books. (2010). ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING AND MARKETING MANAGEMENT. [online] Available at: https://books.google.lk/books?id=iFZwDwAAQBAJ&pg=PA154&lpg=PA154&dq=Albus,+J.+S.+(2002).+%224-D/RCS:+A+Reference+Model+Architecture+for+Intelligent+Unmanned+Ground+Vehicles%22+(PDF&source=bl&ots=-rFrE2Zb0u&sig=ACfU3U1W94025aV5oTVd240IiWXUI33mKA&hl=en&sa=X&ved=2ahUKEwiqoYSyrOnmAhVYfisKHfP0CK4Q6AEwAHoECAoQAQ#v=onepage&q=Albus%2C%20J.%20S.%20(2002).%20%224-D%2FRCS%3A%20A%20Reference%20Model%20Architecture%20for%20Intelligent%20Unmanned%20Ground%20Vehicles%22%20(PDF&f=false [Accessed 4 Jan. 2020].
17. Google Books. (2010). The Quest for Artificial Intelligence. [online] Available at: https://books.google.lk/books?id=nUJdAAAAQBAJ&pg=PT359&lpg=PT359&dq=Brooks,+Rodney+(1990).+%22Elephants+Don%27t+Play+Chess%22+(PDF).+Robotics+and+Autonomous+Systems.&source=bl&ots=2lM5APblLD&sig=ACfU3U2H-vZO0eTnTzYAv-iqQbSem5rrjQ&hl=en&sa=X&ved=2ahUKEwjP7fzTrOnmAhXDWisKHTDPD9oQ6AEwCHoECAkQAQ#v=onepage&q=Brooks%2C%20Rodney%20(1990).%20%22Elephants%20Don’t%20Play%20Chess%22%20(PDF).%20Robotics%20and%20Autonomous%20Systems.&f=false [Accessed 4 Jan. 2020].
18. Buchanan, B.G. (2019). A (Very) Brief History of Artificial Intelligence. AI Magazine, [online] 26(4), pp.53–53. Available at: https://www.aaai.org/ojs/index.php/aimagazine/article/view/1848 [Accessed 4 Jan. 2020].
19. Wikipedia Contributors (2019). Artificial intelligence. [online] Wikipedia. Available at: https://en.wikipedia.org/wiki/Artificial_intelligence [Accessed 4 Jan. 2020].
20. Future of Life Institute. (2015). Benefits & Risks of Artificial Intelligence – Future of Life Institute. [online] Available at: https://futureoflife.org/background/benefits-risks-of-artificial-intelligence/?cn-reloaded=1 [Accessed 4 Jan. 2020].
21. (AI)? (2019). What is Artificial Intelligence (AI)? [online] SearchEnterpriseAI. Available at: https://searchenterpriseai.techtarget.com/definition/AI-Artificial-Intelligence [Accessed 4 Jan. 2020].
22. Sas.com. (2010). Artificial Intelligence – What it is and why it matters. [online] Available at: https://www.sas.com/en_us/insights/analytics/what-is-artificial-intelligence.html [Accessed 4 Jan. 2020].