03_Handout_1(6)emer.pdf

IT2113 Artificial Intelligence intelligence and ability. The concept of artificial superintelligence sees AI evolve to be so akin to human emotions...

IT2113 Artificial Intelligence intelligence and ability. The concept of artificial superintelligence sees AI evolve to be so akin to human emotions and experiences, Artificial Intelligence (AI) is a system that can learn how to learn, or in that it doesn’t just understand them, it evokes emotions, needs, other words a series of instructions (an algorithm) that allows computers to beliefs, and desires of its own. In addition to replicating the multi- write their own algorithms without being explicitly programmed for. faceted intelligence of human beings, ASI would theoretically be exceedingly better at everything we do; math, science, sports, art, There are three levels of artificial intelligence: medicine, hobbies, emotional relationships, everything. ASI would Artificial Narrow Intelligence (ANI) – It refers to AI systems that have a greater memory and a faster ability to process and analyze can only perform a specific task autonomously using human-like data and stimuli. Consequently, the decision-making and problem- capabilities. These machines can do nothing more than what they solving capabilities of super intelligent beings would be far superior are programmed to do, and thus have a very limited or narrow range than those of human beings. of competencies. Artificial narrow intelligence is goal-oriented, designed to perform singular tasks - i.e. facial recognition, speech Types of Artificial Intelligence (AI) recognition/voice assistants, driving a car, or searching the internet - and is very intelligent at completing the specific task it is There are four types of artificial intelligence: programmed to do. While these machines may seem intelligent, Reactive Machines – These are the oldest forms of AI systems they operate under a narrow set of constraints and limitations, that have extremely limited capability. They emulate the human which is why this type is commonly referred to as weak AI. Narrow mind’s ability to respond to different kinds of stimuli. These AI doesn’t mimic or replicate human intelligence, it merely simulates machines do not have memory-based functionality. This means human behavior based on a narrow range of parameters and such machines cannot use previously gained experiences to inform contexts. their present actions, i.e., these machines do not have the ability to Artificial General Intelligence (AGI) – It is also referred to as “learn.” These machines could only be used for automatically strong AI or deep AI. It is the concept of a machine with general responding to a limited set or combination of inputs. They cannot intelligence that mimics human intelligence and/or behaviors, with be used to rely on memory to improve their operations based on the the ability to learn and apply its intelligence to solve any problem. same. A popular example of a reactive AI machine is IBM’s Deep AGI can think, understand, and act in a way that is indistinguishable Blue, a machine that beat chess Grandmaster Garry Kasparov in from that of a human in any given situation. AI researchers and 1997. scientists have not yet achieved strong AI. To succeed, they would Limited Memory – These are machines that, in addition to having need to find a way to make machines conscious, programming a the capabilities of purely reactive machines, are also capable of full set of cognitive abilities. Machines would have to take learning from historical data to make decisions. Nearly all existing experiential learning to the next level, not just improving efficiency applications that we know of come under this category of AI. All on singular tasks, but gaining the ability to apply experiential present-day AI systems, such as those using deep learning, are knowledge to a wider range of different problems. Strong AI uses a trained by large volumes of training data that they store in their theory of mind AI framework, which refers to the ability to discern memory to form a reference model for solving future problems. For needs, emotions, beliefs and thought processes of other intelligent instance, an image recognition AI is trained using thousands of entitles. Theory of mind-level AI is not about replication or pictures and their labels to teach it to name objects it scans. When simulation, it’s about training machines to truly understand humans. an image is scanned by such an AI, it uses the training images as Artificial Super Intelligence – It is an intellect that is much smarter references to understand the contents of the image presented to it, than the best human brain in practically every field, including and based on its “learning experience” it labels new images with scientific creativity, general wisdom, and social skills. ASI is where increasing accuracy. machines become self-aware and surpass the capacity of human 03 Handout 1 *Property of STI  [email protected] Page 1 of 5  IT2113 Theory of Mind – It is the next level of AI systems that researchers AI in Education: Artificial Intelligence can automate grading, giving are currently engaged in innovating. A theory of mind-level AI will educators more time. It can assess students and adapt to their be able to better understand the entities it is interacting with by needs, helping them work at their own pace. AI helps find out what discerning their needs, emotions, beliefs, and thought processes. a student does and does not know, building a personalized study While artificial emotional intelligence is already a budding industry schedule for each learner considering the knowledge gaps. In such and an area of interest for leading AI researchers, achieving theory a way, AI tailors studies according to student’s specific needs, of mind level of AI will require development in other branches of AI increasing their efficiency. AI tutors can provide additional support as well. This is because to truly understand human needs, AI to students, ensuring they stay on track. And it could change where machines will have to perceive humans as individuals whose minds and how students learn, perhaps even replacing some teachers. AI can be shaped by multiple factors, essentially “understanding” systems can use the materials of a traditional syllabus to create humans. customized textbooks for certain subjects. Such systems digitize Self-aware – This is the final stage of AI development which this course material and create new learning interfaces to help currently exists only hypothetically. Self-aware AI, which, self students of all academic grades and ages. In general, AI are helping explanatorily, is an AI that has evolved to be so akin to the human to eliminate boundaries and extend educational opportunities to brain that it has developed self-awareness. This type of AI will not learners throughout the world. only be able to understand and evoke emotions in those it interacts AI in Finance: Artificial Intelligence and the finance industry are a with, but also have emotions, needs, beliefs, and potentially desires match made in heaven. The financial sector relies on accuracy, of its own. This is because once self-aware, the AI would be capable real-time reporting and processing high volumes of quantitative of having ideas like self-preservation which may directly or indirectly data to make decisions, all areas intelligent machines excel in. As spell the end for humanity, as such an entity could easily the industry takes note of AI's efficiency and accuracy, it is rapidly outmaneuver the intellect of any human being and plot elaborate implementing automation, chatbots, adaptive intelligence, schemes to take over humanity. algorithmic trading and machine learning into financial processes. One of the biggest financial trends is the robo-advisor, an Applications of Artificial Intelligence automated portfolio manager. These automated advisors use AI and algorithms to scan data in the markets and predict the best Artificial Intelligence has various applications in today’s society. It is stock or portfolio based on preferences. becoming essential for today’s time because it can solve complex problems AI in Marketing: Artificial intelligence applications are popular in in an efficient way in multiple industries. The following are some sectors that the marketing domain as well. Artificial intelligence is helping have the application of Artificial Intelligence: marketers build in-depth customer insight reports, power pertinent AI in Healthcare: Artificial intelligence finds diverse applications in content creation and book more impactful business meetings – all the healthcare sector. AI is used in healthcare to build sophisticated without a large human influence. Using AI, marketers can deliver machines that can detect diseases and identify cancer cells. AI can highly targeted and personalized ads with the help of behavioral help analyze chronic conditions with lab and other medical data to analysis, pattern recognition, etc. It also helps with retargeting ensure early diagnosis. AI uses the combination of historical data audiences at the right time to ensure better results and reduced and medical intelligence for the discovery of new drugs. The biggest feelings of distrust and annoyance. AI can help with content bets are on improving patient outcomes and reducing costs. The marketing in a way that matches the brand's style and voice. It can role of AI in healthcare enables machines to interpret the medical be used to handle routine tasks like performance, campaign history of a patient and predict possible diseases that the individual reports, and much more. AI can provide users with real-time can become vulnerable to in the coming years. One of the best- personalization based on their behavior and can be used to edit and known healthcare technologies is IBM Watson. It understands optimize marketing campaigns to fit a local market's needs. natural language and can respond to questions asked of it. 03 Handout 1 *Property of STI  [email protected] Page 2 of 5  IT2113 AI in Social Media: Ever since social media has become our AI in Travel and Transportation: Artificial intelligence is becoming identity, we’ve been generating an immeasurable amount of data a mega-trend in the travel and transportation industries. From through chats, tweets, posts, and so on. And wherever there is an making travel arrangements to suggesting the most efficient route abundance of data, AI is always present. In social media platforms home after work, AI is making it easier to get around. Travel like Facebook, AI is used for face verification wherein machine companies are especially capitalizing on ubiquitous smartphone learning and deep learning concepts are used to detect facial usage. A majority of users claim they book trips on their phones, features and tag your friends. Deep Learning is used to extract review travel tips, and research local landmarks and restaurants. every minute detail from an image by using a bunch of deep neural AI-powered chatbots are rapidly changing the travel industry by networks. Another such example is Twitter’s AI, which is being used facilitating human-like interaction with customers for faster to identify hate speech and terroristic language in tweets. It makes response times, better booking prices, and even travel use of Machine Learning, Deep Learning, and Natural language recommendations. For example, telling a travel chatbot you want to processing to filter out offensive content. On Instagram, AI go to Paris might yield a natural language response suggesting considers your likes and the accounts you follow to determine what flights, hotels, and things to do in the City of Light based on a user's posts you are shown on your explore tab. preferences culled from the conversation. AI in E-Commerce: Ever scrolled through a website only to find an AI in Robotics: Robotics is another field where artificial intelligence image of the exact shirt you were just looking at on another site pop applications are commonly used. It is a well-known fact that artificial up again? You can thank artificial intelligence for that. Artificial intelligence is the driving force behind robotics. Usually, general Intelligence technology is used to create recommendation engines robots are programmed such that they can perform some repetitive through which you can engage better with your customers. These tasks, but with the help of AI, we can create intelligent robots which recommendations are made in accordance with their browsing can perform tasks with their own experiences without being pre- history, preference, and interests. It helps in improving your programmed. The use of artificial intelligence in robotics has made relationship with your customers and their loyalty to your brand. possible the presence of robots in multiple industries like finance, Companies use artificial intelligence to deploy chatbots, predict marketing, and healthcare. A real-life example of this application is purchases and gather data to create a more customer-centric e- the robot - Sophia. Considered to be a 'humanoid', Robot Sophia is commerce experience. Credit card frauds and fake reviews are two a blend of a human and a robot with in-built abilities of both. A robot of the most significant issues that E-Commerce companies deal powered by AI uses real-time updates to sense obstacles in their with. By considering the usage patterns, AI can help reduce the path and pre-plan its journey instantly. possibility of credit card frauds taking place. AI in Space Exploration: Artificial intelligence is not only present AI in Automated Cars: Artificial Intelligence is the way how self- on Earth but in outer space too. When it comes to the applications driving cars work. Automated cars are a typical example of of AI, machines have traversed outer space and led humans to applications of AI that explains the advancements of the field along explore outer space. Space expeditions and discoveries always with the automobile industry. The thought of cars driving by require analyzing vast amounts of data. Artificial intelligence is the themselves with algorithms of artificial intelligence is the future of best way to handle and process data on this scale. Artificial the automobile industry. These cars are loaded with sensors that intelligence is also being used for NASA’s next rover mission to are constantly taking note of everything going on around the car Mars, the Mars 2020 Rover. The rover is responsible for and using AI to make the correct adjustments. These sensors autonomous targeting of cameras in order to perform investigations capture thousands of data points every millisecond (like car speed, on Mars. Whether it is about Mars missions or satellite installments road conditions, pedestrian whereabouts, other traffic, etc.), and in the exosphere, AI is persistently heading towards the exploration use AI to help interpret the data and act accordingly – all in a blink- of outer space. Some of the aspects that involve the application of of-an-eye. Big corporations like Tesla have been working on artificial intelligence in outer space exploration are map-building, automated cars and are heading towards this development. satellite navigation, and location tracking technology. 03 Handout 1 *Property of STI  [email protected] Page 3 of 5  IT2113 Machine Learning Concepts Autonomic Computing Machine Learning is an application of artificial intelligence that provides Autonomic Computing (AC) has been inspired by the human autonomic systems the ability to automatically learn and improve from experience nervous system and is used to manage such complex and sophisticated without being explicitly programmed. Machine learning focuses on the systems. Just as the human body acts and responds without the individual development of computer programs that can access data and use it to learn controlling functions (e.g., internal temperature rises and falls, breathing for themselves. The primary aim is to allow the computers learn rate fluctuates, glands secrete hormones in response to stimulus), the automatically without human intervention or assistance and adjust actions autonomic computing environment operates organically in response to the accordingly. Machine learning algorithms are often categorized as: input it collects. Autonomic computing shared the vision of making all Supervised Machine Learning – It applies what has been learned computing systems manage themselves automatically. It refers to self- in the past to new data using labeled examples to predict future managing characteristics of distributed computing resources, which events. Starting from the analysis of a known training dataset, the recognize and understand changes in the system, take appropriate learning algorithm produces an inferred function to make corrective actions completely automatically, with close to zero human predictions about the output values. The system is able to provide intervention. Characteristics of autonomic computing-based systems are: targets for any new input after sufficient training. The learning Self-Configuring – Using this property, the autonomic system must algorithm can also compare its output with the correct, intended adjust and adapt itself in changing environments automatically output and find errors in order to modify the model accordingly. using the policies provided by IT professionals for meeting defined Unsupervised Machine Learning – In contrast to supervised business goals. Self-configuring provides maintainability, learning, these are used when the information used to train is functionality, portability, and usability to the system. An AC system neither classified nor labeled. Unsupervised learning studies how should be able to configure and reconfigure itself under varying and systems can infer a function to describe a hidden structure from unpredictable conditions. System configuration or setup must occur unlabeled data. The system doesn’t figure out the right output, but automatically, as well as dynamic adjustments to the configuration it explores the data and can draw inferences from datasets to to best handle a changing operating environment. Such dynamic describe hidden structures from unlabeled data. adjustments may also mean the addition of new hardware Semi-supervised Machine Learning – It falls somewhere in resources in response to management software or policies defined between supervised and unsupervised learning, since they use by a systems administrator. both labeled and unlabeled data for training – typically a small Self-Healing – Using the self-healing characteristic, it can heal amount of labeled data and a large amount of unlabeled data. The itself and its components. It identifies fault, repair, or replace the systems that use this method are able to considerably improve components automatically. An AC system should be able to detect learning accuracy. Usually, semi-supervised learning is chosen and recover from potential problems and continue to function when the acquired labeled data requires skilled and relevant smoothly. Self-healing components identify the malfunctioning and resources in order to train it / learn from it. Otherwise, acquiring then take corrective action without disrupting the IT environment. unlabeled data generally doesn’t require additional resources. Self-healing can be done in two modes – reactive mode and Reinforcement Machine Learning – This is a learning method that proactive mode. In reactive mode, when fault occurs, healing interacts with its environment by producing actions and discovers components react accordingly and try to recover from it. In proactive errors or rewards. Trial and error search and delayed reward are mode, the system continuously monitors and tries to prevent the the most relevant characteristics of reinforcement learning. This system from faults or any disruption. The main objective of self- method allows machines and software agents to automatically healing is to maximize the availability, maintainability, survivability, determine the ideal behavior within a specific context in order to and reliability of the autonomic application and system. In short, maximize its performance. self-healing is the ability to recover from external damage or internal errors. 03 Handout 1 *Property of STI  [email protected] Page 4 of 5  IT2113 Self-Optimizing – Through the self-optimizing characteristic, an autonomic software system will try to be more efficient, tune its components dynamically, improve its performance and execution, and the system will find optimal ways for improving overall utilization. It provides efficiency, maintainability, and functionality to autonomous systems. An AC system should be able to detect suboptimal behaviors and optimize itself to improve its execution. It should be “goal” oriented, i.e. it should pro-actively look for opportunities to optimize its use. In short, self-optimizing is the ability to manage all resources and components to optimize operation. Self-Protecting – It is the ability to combat external threats to operations. An AC system should be capable of detecting and protecting its resources from both internal and external attacks and maintaining overall system security and integrity. Self-protecting provides reliability and functionality to the systems. These systems must be able to detect hostile behaviors and other problems from the reports generated by sensors and must be able to defend them. Hostile behaviors can be virus attacks, accidental attacks, malicious attacks, unauthorized access, system failure and denial of service attacks, etc. ___________________________________________________________ References: Applications of artificial intelligence (2021). Citing sources. Retrieved on September 14, 2021, from https://builtin.com/ Levels of artificial intelligence (2019). Citing sources. Retrieved on September 14, 2021, from https://www.forbes.com/ Low, A. & Lawless, S., (2021). Artificial intelligence foundations. BCS Learning and Development Ltd. Machine learning concepts (2021). Citing sources. Retrieved on September 14, 2021, from https://www.expert.ai/ Overview of autonomic computing systems (2017). Citing sources. Retrieved on September 14, 2021, from https://www.ijser.org/ Types of artificial intelligence (2021). Citing sources. Retrieved on September 14, 2021, from https://codebots.com/ 03 Handout 1 *Property of STI  [email protected] Page 5 of 5