CS3014: Artificial Intelligence INTRODUCTION TO ARTIFICIAL INTELLIGENCE

Philip Apodo Oyier oyier@itc.jkuat.ac.ke BIT 2319 Lecture 2: Introduction to Artificial Intelligence.

Learning Outcomes At the end of this course:  Knowledge and understanding You should have a knowledge and understanding of the basic concepts of Artificial Intelligence including Search, Game Playing, KBS (including Uncertainty), Planning and Machine Learning.  Intellectual skills You should be able to use this knowledge and understanding of appropriate principles and guidelines to synthesise solutions to tasks in AI and to critically evaluate alternatives.  Practical skills You should be able to use a well known declarative language (Prolog, Jess, Clips, Python) and to construct simple AI systems.  Transferable Skills You should be able to solve problems and evaluate outcomes and alternatives 2.

Information Retention Based on the Level of Active Learner Involvement Learning Pyramid.

Learning Pyramid: Information Retention Based on the Level of Active Learner Involvement Source : Adopted from National Training Laboratories. Institutes of Applied Behavioral Science 4.

Analytic and Global Learning Theory  This theory describes the order in which a learner prefers to process information received by looking at the whole then breaking it down into individual parts or by looking at each individual part and then combining it into a whole.  Sometimes called right-brain and left-brain. 5.

Global Learner (Right-Brain)  Needs to process the big picture (overall) view first then can concentrate on the individual parts that make up the big picture.  They are uncomfortable learning when they do not have a sense of the big picture.  These students appreciate an overview of the material before you start teaching.  Process information globally and simultaneously, deals in images.  Tend to be creative, artistic, imaginative, emotional, and intuitive and generally like working on teams.  Try mental imagery, drawing, maps, metaphors, music and dance, experiential learning. 6.

Analytic Learner (Left-Brain)  Process information logically, sequentially, in small parts.  They are uncomfortable with learning that is occurring out of sequence.  Tend to enjoy spelling, numbers, thinking, reading, analysis and speaking.  Try lectures with outlines, reading assignments, and multiple-choice exams. 7.

Differences Between Analytic and Global Learners The forest or the trees?  Analytic Learners - separate the forest from the trees: analytic learners look at every tree in the forest before being comfortable enough to declare that they are in the forest.  Global Learners - will walk up to several trees, quickly declare it is a forest, and then will begin to look at the individual trees. 8.

Intelligence.

Intelligence  Intelligence  “The capacity to learn and solve problems” (Websters dictionary)  “The ability to comprehend; to understand and profit from experience”  Capacity of mind, especially to understand principles, truths, facts or meanings, acquire knowledge, and apply it to practice  The ability to apply knowledge to manipulate one's environment or to think abstractly as measured by objective criteria (as tests) 10.

Theory of Multiple Intelligence's  A. Howard Gardner described “multiple intelligence's”.  Gardner’s idea was that measuring “Intelligence Quotient (IQ), a measure of your ability to reason and solve problems, through a series of cognitive exercises does not fully measure the range of intelligences expressed by each individual.  Gardner’s Multiple Intelligence's  Hypothesized that each person has aptitude (a natural ability to do something) in the following areas, with each individual having some areas with greater aptitude than others. 11.

Types of Intelligence (Multiple Intelligence theory, Howard Gardner)  The theory of multiple intelligences was proposed by Howard Gardner in 1983 to explore and articulate various forms or expressions of intelligence available to cognition  Howard Gardner is a psychologist and Professor at Harvard University's Graduate School of Education.  Based on his study of many people from many different walks of life in everyday circumstances and professions, Gardner developed the theory of multiple intelligences.  Gardner defined the first seven intelligences in Frames of Mind in 1983. He added the last two in Intelligence Reframed in 1999. 12.

Types of Intelligence (Multiple Intelligence Theory, Howard Gardner) Intelligence Skills and Career Preferences Verbal/Linguistic Intelligence Well-developed verbal skills and sensitivity to the sounds, meanings and rhythms of words Skills - Listening, speaking, writing, teaching. Careers - Poet, journalist, writer, teacher, lawyer, politician, translator Mathematical/Logical Intelligence Ability to think conceptually and abstractly, and capacity to discern logical or numerical patterns Skills - Problem solving (logical & math), performing experiments Careers - Scientists, engineers, accountants, mathematicians Musical/Rhythmic Intelligence Ability to produce and appreciate rhythm, pitch and timber Skills - Singing, playing instruments, composing music Careers - Musician, disc jockey, singer, composer Visual/Spatial Intelligence Capacity to think in images and pictures, to visualize accurately and abstractly Skills - puzzle building, painting, constructing, fixing, designing objects Careers - Sculptor, artist, inventor, architect, mechanic, engineer 13.

Types of Intelligence (Multiple Intelligence theory, Howard Gardner) Intelligence Skills and Career Preferences Bodily/Kinesthetic Intelligence Ability to control one's body movements and to handle objects skillfull Skills - Dancing, sports, hands on experiments, acting Careers - Athlete, PE teacher, dancer, actor, firefighter Interpersonal Intelligence Capacity to detect and respond appropriately to the moods, motivations and desires of others Skills - Seeing from other perspectives, empathy, counseling, co-operating Careers - Counselor, salesperson, politician, business person, minister Intrapersonal Intelligence Capacity to be self-aware and in tune with inner feelings, values, beliefs and thinking processes Skills - Recognize one’s S/W, reflective, aware of inner feelings Careers - Researchers, theorists, philosophers Naturalist Intelligence Ability to recognize and categorize plants, animals and other objects in nature Skills - Recognize one’s connection to nature, apply science theory to life Careers – Scientist, naturalist, landscape architect 14.

Types of Intelligence (Multiple Intelligence theory, Howard Gardner) Intelligence Skills and Career Preferences Existential Intelligence Sensitivity and capacity to tackle deep questions about human existence, such as the meaning of life, why do we die, and how did we get here Skills – Reflective and deep thinking, design abstract theories Careers – Scientist, philosopher, theologian 15 Note: Understanding the various types of intelligence provides theoretical foundations for recognizing different talents and abilities in people. The Nature of Intelligence Characteristics of intelligent behavior include the ability to: • Learn from experience and apply knowledge acquired from experience • Handle complex situations • Solve problems when important information is missing • Determine what is important • React quickly and correctly to a new situation.

Artificial Intelligence.

AI in the Eyes of the Society 17  People get to know AI through news, movies, and actual applications in daily life. What is AI in the eyes of the public? The Terminator 2001: A Space Odyssey The Matrix I, Robot Blade Runner Elle Bicentennial Man … Self-service security check Spoken language evaluation Music/Movie recommendation Smart speaker … News AI Applications AI industry outlook Challenges faced by AI … Movies AI Control over human beings Fall in love with AI Self-awareness of AI … Applications in daily life Security protection Entertainment Smart Home Finance ….

The Rise of AI  In March 2016, AlphaGo defeated Lee Sedol, a South Korean 9-dan professional Go player, by 4-1. This reshaped people's opinion on AI and unveiled its overwhelming development.  AlphaGo:  On January 27, 2016, Nature, an international top journal, reported that AlphaGo, a computer Go program developed by Google, defeated Fan Hui, a 2-dan professional player who once won the European Professional Go Championship by 5-0 without handicap.  This was an unprecedented breakthrough in the AI field of Go matches. It was also the first time that a computer Go program had beaten a professional human player on a full- sized board without handicap. 18.

Dartmouth Workshop: Birth of AI  In August 1956, some scientists and mathematicians gathered at Dartmouth College, discussing about how to make machines simulate human learning and any other feature of intelligence.  They were John McCarthy (creator of the Lisp programming language), Marvin Minsky (AI and cognitive scientist), Claude Shannon (father of information theory), Allen Newell (computer scientist), and Herbert A. Simon (winner of the Nobel Prize in Economic Sciences).  The workshop ran for two months. No consensus was reached, but they picked the name artificial intelligence for the field they discussed about. Then, the year 1956 marked the birth of AI. 19.

AI Development History Summary 20 1956-1976 Flourishing period The Dartmouth Workshop coined the concept and goals of Artificial Intelligence. 1950s 1960s 1970s 1980s 1990s 2000s 2010s 2020s 1976-1982 Declining period Artificial Intelligence was criticized due to its insufficient computing capability, computational complexity, and difficulties in common sense based reasoning. 1982-1987 Flourishing period Expert systems capable of logical deduction and solving problems in specific domains prevailed and fifth generation computers showed up. 1987-1997 Declining period The technical field encountered its bottleneck again, abstract reasoning lost its popularity, and models based on symbolic processing lost their advantages. 1997-2010 Recovering period The computing performance was improved and the Internet technologies were widely used. 2010- Explosive growth period Cutting-edge information technologies lead to the transformation of information environment and data. The emergence of multiple-model data such as massive images, voices, and texts improves the computing capability. The concept of Artificial Intelligence was coined at the Dartmouth Workshop in 1956. In 1959, Arthur Samuel put forward the concept of machine learning. In 1976, the failure of some projects such as machine translation and negative effects of some academic reports led to reduction of artificial intelligence expenditure. In 1985, decision tree models with stronger visible effects and multi- layer artificial neural networks that broke through limitations of early sensing machines showed up. In 1987, the LISP market collapsed. In 1997, Deep Blue defeated Garry Kasparov, a world chess champion. In 2006, Geoffrey Hinton and his students initiated the study on deep learning. In 2010, the big data era came. In 2014, Microsoft released the world's first virtual assistant, Cortana. In March 2016, AlphaGo won the world chess championship Lee Sedol by 4:1. In October 2017, the Deep Mind team released AlphaGo Zero, the strongest version..

AI in the Eyes of Researchers  "I propose to consider the question, 'Can machines think?'“-Alan Turing 1950  The branch of computer science concerned with making computers behave like humans- John McCarthy 1956  The science of making machines do things that would require intelligence if done by men - Marvin Minsky 21.

Mainstream AI Theories.

Symbolicism  Symbolicism (Logicism, psychologism, computerism)  Principle: Physical symbol system hypothesis and finite reasonableness principle  Origin: Mathematical logic  Concept:  Symbol is the human cognition unit, and the cognition process is a symbol operation process.  People are regarded as a physical symbol system, so are computers. Therefore, computers can be used to simulate human behavior.  Knowledge is a form of information and is the basis of intelligence. The critical issues of AI are knowledge representation and knowledge inference.  Representatives: Allen Newell, Herbert Alexander Simon, Nilsson, etc. 23.

Symbolicism Contd. 24 Symbolicism Representatives.

Connectionism  Connectionism  Principle: neural network, connection mechanism and learning algorithm between neural networks  Origin: bionics, especially the study of the human brain model  Concept:  Neuron, instead of the symbol operation process, is the basic thinking unit.  Human brain differs from computers, and the human brain pattern can be used to replace the computer pattern.  Representatives: Warren McCulloch, Walter Pitts, John Hopfield, Rumelhart, D.E., etc. 25.

Connectionism Contd. 26 Connectionism Representatives Syna pse Syna pse Axon Dend rite Pregang lionic neuron axon Cytoplast Nucleus Syna pse Next neuron dendrite.

Actionism  Actionism (evolutionism and cyberneticsism)  Principle: cybernetics and perception-action control system  Origin: cybernetics  Concept:  Intelligence depends on perception and actions. The "perception-action" mode of intelligent behavior is proposed.  Intelligence requires no knowledge, representation, and inference. Artificial intelligence can evolve like human intelligence. Intelligent behavior can only interact with the surrounding environment in the real world. 27.

What Is Artificial Intelligence?  Artificial Intelligence (AI) is a technical science that studies and develops theories, methods, technologies, and applications for simulating and extending human intelligence. This term was first coined by John McCarthy in 1956.  McCarthy defined the subject as the "science and engineering of making intelligent machines, especially intelligent computer programs".  The purpose of AI is to enable machines to think like people and to make machines intelligent.  Today, AI has become an interdisciplinary course that involves various fields. 28 AI Brain science Cognitive science Psychology Linguistics Logic Philosophy Computer Science.

Relationship of AI, Machine Learning, and Deep Learning 29.

Relationship of AI, Machine Learning and Deep Learning  AI: A new technical science that focuses on the research and development of theories, methods, techniques, and application systems for simulating and extending human intelligence.  Machine learning: A core research field of AI. It focuses on the study of how computers can obtain new knowledge or skills by simulating or performing learning behavior of human beings, and reorganize existing knowledge architecture to improve its performance. It is one of the core research fields of AI.  Deep learning: A new field of machine learning. The concept of deep learning originates from the research on artificial neural networks. The multi-layer perceptron (MLP) is a type a deep learning architecture. Deep learning aims to simulate the human brain to interpret data such as images, sounds, and texts. 30.

Artificial Intelligence Approaches Systems that act rationally Systems that think like humans Systems that think rationally Systems that act like humans THOUGHT BEHAVIOUR HUMAN RATIONAL 31.

Four main approaches to AI  Systems that act like humans  AI is: “The art of creating machines that perform functions that require intelligence when performed by people” (Kurzweil)  Ultimately to be tested by the Turing Test  Systems that think like humans  AI is: “[The automation of] activities that we associate with human thinking, activities such as decision- making, problem solving, learning…” (Bellman)  Goal is to build systems that function internally in some way similar to human mind 32.

Four main approaches to AI Contd.  Systems that think rationally  AI is: “The study of the computations that make it possible to perceive, reason, and act” (Winston)  Approach firmly grounded in logic i.e., how can knowledge be represented logically, and how can a system draw deductions?  Uncertain knowledge? Informal knowledge? “I think I love you.”  Systems that act rationally  AI is: “The branch of computer science that is concerned with the automation of intelligent behavior” (Luger and Stubblefield)  The intelligent agent approach  An agent is something that perceives and acts  Emphasis is on behavior 33.

Systems that act like humans: Turing Test  “The art of creating machines that perform functions that require intelligence when performed by people.” (Kurzweil)  “The study of how to make computers do things at which, at the moment, people are better.” (Rich and Knight)  Computational models of human behaviour  Programs that behave (externally) like humans.  This is the original idea from Turing and the well known Turing Test is to use to verify this 34.

Systems that Act Like Humans: Turing Test  You enter a room which has a computer terminal. You have a fixed period of time to type what you want into the terminal, and study the replies. At the other end of the line is either a human being or a computer system.  If it is a computer system, and at the end of the period you cannot reliably determine whether it is a system or a human, then the system is deemed to be intelligent. 35.

Systems that Act like humans  The Turing Test approach  A human questioner cannot tell if  There is a computer or a human answering his question, via teletype (remote communication)  The computer must behave intelligently  Intelligent behavior  To achieve human-level performance in all cognitive tasks 36.

Systems that act like humans  What means “behave rationally” for a person/system:  Take the right/ best action to achieve the goals, based on his/its knowledge and belief  Example. Assume I don’t like to get wet (my goal), so I bring an umbrella (my action). Do I behave rationally?  The answer is dependent on my knowledge and belief  If I’ve heard the forecast for rain and I believe it, then bringing the umbrella is rational.  If I’ve not heard the forecast for rain and I do not believe that it is going to rain, then bringing the umbrella is not rational 37.

Systems that act like humans  Note on behave rationally or rationality  “Behave rationally” does not always achieve the goals successfully  Example.  My goals – (1) do not get wet if rain; (2) do not be looked stupid (such as bring an umbrella when no raining)  My knowledge/belief – weather forecast for rain and I believe it  My rational behaviour – bring an umbrella  The outcome of my behaviour: If rain, then my rational behaviour achieves both goals; If not rain, then my rational behaviour fails to achieve the 2nd goal  The successfulness of “behave rationally” is limited by my knowledge and belief 38.

Systems that act like humans  Note on behave rationally or rationality  Another limitation of “behave rationally” is the ability to compute/ find the best action  In chess-playing, it is sometimes impossible to find the best action among all possible actions  So, what we can really achieve in AI is the limited rationality  Acting based to your best knowledge/belief (best guess sometimes)  Acting in the best way you can subject to the computational constraints that you have 39.

Systems that act like humans  These cognitive tasks include:  Natural language processing  for communication with human  Knowledge representation  to store information effectively & efficiently  Automated reasoning  to retrieve & answer questions using the stored information  Machine learning  to adapt to new circumstances 40.

The total Turing Test  Includes two more issues:  Computer vision  to perceive objects (seeing)  Robotics  to move objects (acting) 41.

What is Artificial Intelligence ? Systems that act rationally Systems that think like humans Systems that think rationally Systems that act like humans THOUGHT BEHAVIOUR HUMAN RATIONAL 42.

Systems that think like humans: cognitive modeling  Humans as observed from ‘inside’  How do we know how humans think?  Introspection vs. psychological experiments  Cognitive Science  “The exciting new effort to make computers think … machines with minds in the full and literal sense” (Haugeland)  “[The automation of] activities that we associate with human thinking, activities such as decision-making, problem solving, learning …” (Bellman) 43.

What is Artificial Intelligence ? Systems that act rationally Systems that think like humans Systems that think rationally Systems that act like humans THOUGHT BEHAVIOUR HUMAN RATIONAL 44.

Systems That Think ‘Rationally’ “Laws of Thought"  Humans are not always ‘rational’  Rational - defined in terms of logic?  Logic can’t express everything (e.g. uncertainty)  Logical approach is often not feasible in terms of computation time (needs ‘guidance’)  “The study of mental facilities through the use of computational models” (Charniak and McDermott)  “The study of the computations that make it possible to perceive, reason, and act” (Winston) 45.

What is Artificial Intelligence ? Systems that act rationally Systems that think like humans Systems that think rationally Systems that act like humans THOUGHT BEHAVIOUR HUMAN RATIONAL 46.

Systems that Act Rationally: “Rational agent”  Rational behavior: doing the right thing  The right thing: that which is expected to maximize goal achievement, given the available information  Giving answers to questions is ‘acting’  I don't care whether a system:  Replicates human thought processes  Makes the same decisions as humans  Uses purely logical reasoning 47.

Systems that Act Rationally  Logic  only part of a rational agent, not all of rationality  Sometimes logic cannot reason a correct conclusion  At that time, some specific (in domain) human knowledge or information is used  Thus, it covers more generally different situations of problems  Compensate the incorrectly reasoned conclusion 48.

Systems that Act Rationally  Study AI as rational agent  Two advantages  It is more general than using logic only  Because: LOGIC + Domain knowledge  It allows extension of the approach with more scientific methodologies 49.

Rational agents  An agent is an entity that perceives and acts  This course is about designing rational agents  Abstractly, an agent is a function from percept histories to actions: [f: P* A]  For any given class of environments and tasks, we seek the agent (or class of agents) with the best performance  Caveat: computational limitations make perfect rationality unachievable  design best program for given machine resources 50.