Unit 07 - Artificial Intelligence PDF

DS Unit 07: Artificial Intelligence Intelligence Intelligence is the ability to learn about, learn from, understand, and interact with one’s environment. Some of the specific abilities that point towards the presence of intelligence are: Adaptability to a new environment or to changes in th...

DS Unit 07: Artificial Intelligence Intelligence Intelligence is the ability to learn about, learn from, understand, and interact with one’s environment. Some of the specific abilities that point towards the presence of intelligence are: Adaptability to a new environment or to changes in the current environment. Capacity for knowledge and the ability to acquire it. Capacity for reason and abstract thought. Ability to comprehend relationships. Ability to evaluate and judge. Capacity for original and productive thought. AI (Artificial Intelligence) The effort to develop the intelligence in the machine, in other words «artificial intelligence» has immensely increased in recent years. AI can be defined as the idea that a computer can imitate and go beyond human intelligence and automatically perform tasks without human involvement. Types of AI The following list is a compilation of AI types. These are not the only types, just the most popular ones. Strong (or Full, or General) AI Weak (or Narrow) AI Domain-specified AI Strong AI (or Full AI, or General AI) is indistinguishable from the human mind. But just like a child, needs to learn through input and experiences, constantly progressing and advancing its abilities and goals over time. Examples of strong AI so far only exist in sci-fi movies such as 2001: A Space Odyssey, WALL-E, and Her. Strong AI can learn new skills it wasn't initially programmed for and develop goals independently. Weak AI (or Narrow AI) is a type of artificial intelligence that is limited to a its own programming. It can perform the tasks that it is programmed for. Since it is AI, it can learn from data input, but cannot create new programming for itself. Siri, Alexa, Deep Blue (chess AI) are some examples of weak AI. Domain-specific AI can automate tasks that humans do, but the focus isn't to imitate humans. It won't be able to switch from one task to another and then dive into a conversation about its plans for the weekend. However, in certain well-defined tasks, domain-specific AI greatly outperforms humans. Person or Machine: The Turing Test With the help of AI, a computer system may attempt to operate as if it is a human being. Thus there is a necessity for a system to tell computers and humans apart. The Turing Test is considered to be the first test to distinguish between humans and machines. The Turing Test is a test devised by the English mathematician Alan M. Turing in 1950, to determine whether or not a computer can be said to think like a human brain. The Test: In order to perform the test, an interrogator and two other players (one of which is a computer system) sit in isolated rooms. The interrogator asks text based questions to the two players, just like text chatting. If the interrogator cannot distinguish the computer system from human, the said computer will have passed the test. AI Techniques: Machine Learning Machine learning is a scientific discipline that explores the construction and study of algorithms that can learn from data. Such algorithms operate by building a model from example inputs and using that to make predictions or decisions, rather than following strictly static program instructions. Supervised Learning vs. Unsupervised Learning vs. Reinforced Learning A machine may learn from samples that are pre-labeled (supervised learning); or by reasoning, using samples that are totally unlabeled or unknown (unsupervised learning). Just like a child to whom a parent teaches what apple is vs. an animal that samples the apple without supervision. Reinforced learning is when a machine interacts with its environment, performs actions, and learns by a trial-and-error method. AI Techniques: Deep Learning Deep Learning is an advanced form of Machine Learning that utilizes artificial neural networks. An artificial neural network is a computing system designed to simulate the way the human brain analyzes and processes information. For example: In image processing, Machine Learning might be able to highlight the edges of an image, while Deep Learning is able to recognize faces. Uses of AI Pattern Recognition Facial Recognition Speech Recognition & Natural Language Processing Gait (walking pattern) Recognition Image Analysis & Recognition (e.g. Google Image Search) Areas of Application Today A.I. is utilized in: Web search engines  Language translation Chess Voice recognition Video games Predictive text Natural language communication Pattern recognition: OCR, image analysis, speech recognition Uses of Artificial Neural Networks Artificial neural networks can learn and model complex and non-linear relationships and can generalize from initial inputs. Facial recognition Forecasting (weather, stock exchange etc.) Music composition are some areas where artificial neural networks are used. AI Winters In the history of AI, there were time periods when interest and funding for AI drastically decreased. These periods are referred to as AI winters. Two major AI winters were in 1974–1980 and 1987–1993. In that sense, since 2005 the world is in an “AI summer”. Singularity and Multiplicity Some scientists predict that the future of AI is going to be a singularity. Some others claim that we will rather witness a multiplicity. Singularity refers to the moment where AI is no longer under human control. Multiplicity refers to a future where humans and machines can coexist where diverse combinations of people and machines work together to solve problems and innovate. AI Dilemmas Following considerations are among the most prominent issues about the design of AI: Fairness and bias in design and use Accountability in design and use Transparency in design and use Uneven and underdeveloped laws, regulations and governance Automation and displacement of humans in multiple contexts and roles