2Jimma University,JiT Depatment of Computing Introduction To Artificial Intelligence Zelalem H.
3Outline (1)Introduction (2) Intelligent Agents What is AI? Foundations of AIState of the art in AIHistory of AI (Reading assignment)(2) Intelligent AgentsAgents and environmentsRationalityThe Nature of environmentsThe structure of agents
41. IntroductionFor thousands of years, we have tried to understand how we think! AI, goes further still; it attempts not just to understand but also to build intelligent entities.
5Introduction… What is AI? Some possible definitions Thinking humanly Thinking rationallyActing humanly Acting rationally
6Introduction… Thinking humanly Cognitive science: the brain as an information processing machineRequires scientific theories of how the brain worksHow to understand cognition as a computational process?Introspection: try to think about how we thinkPredict and test behavior of human subjectsImage the brain, examine neurological data
7Introduction… Acting humanly The Turing Test What capabilities would a computer need to have to pass the Turing Test?Natural language processingKnowledge representationAutomated reasoningMachine learning
8Introduction… Turing Test: Criticism What are some potential problems with the Turing Test?Some human behavior is not intelligentSome intelligent behavior may not be humanHuman observers may be easy to foolChinese room argument: one may simulate intelligence without having true intelligenceIs passing the Turing test a good scientific goal?
9Introduction… Thinking rationally Idealized or “right” way of thinking Logic: patterns of argument that always yield correct conclusions when supplied with correct premises“Socrates is a man;All men are mortal;Therefore Socrates is mortal.”Beginning with Aristotle, philosophers and mathematicians have attempted to formalize the rules of logical thought
10Introduction… Thinking rationally … Logicist approach to AI: describe problem in formal logical notation and apply general deduction procedures to solve itProblems with the logicist approachComputational complexity of finding the solutionDescribing real-world problems and knowledge in logical notationA lot of intelligent or “rational” behavior has nothing to do with logic
11Introduction… Acting rationally A rational agent is one that acts to achieve the best outcomeGoals are application-dependent and are expressed in terms of the utility of outcomesBeing rational means maximizing your expected utilityThis definition of rationality only concerns the decisions/actions that are made, not the cognitive process behind them
12Introduction… Acting rationally… Any disadvantages? Advantages Generality: goes beyond explicit reasoning, and even human cognition altogetherPracticality: can be adapted to many real-world problemsAmenable to good scientific and engineering methodologyAvoids philosophy and psychologyAny disadvantages?Not feasible in complicated envt’sComputational demands are just too high
13AI Foundations Philosophy Mathematics Can formal rules be used to draw valid conclusions?How does the mind arise from a physical brain?Where does knowledge come from?How does knowledge lead to action?MathematicsWhat are the formal rules to draw valid conclusions?What can be computed?How do we reason with uncertain information?
14AI Foundations… Economics Neuroscience Psychology How should we make decisions so as to maximize payoff?How should we do this when others may not go along?NeuroscienceHow do brains process information?PsychologyHow do humans and animals think and act?
15AI Foundations… Computer Engineering How can we build an efficient computer?Control Theory and CyberneticsHow can artifacts operate under their own control?LinguisticsHow does language relate to thought?
17The History of Artificial intelligence Reading AssignmentFrom Page 16 to 28Russell and Norvig, 3rd edition
182. Intelligent AgentsAn agent is anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators
19Agent functionThe agent function maps from percept histories to actionsThe agent program runs on the physical architecture to produce the agent functionagent = architecture + program
20Vacuum-cleaner world Percepts: Location and status, e.g., [A,Dirty] Actions: Left, Right, Suck, NoOpfunction Vacuum-Agent([location,status]) returns an actionif status = Dirty then return Suckelse if location = A then return Rightelse if location = B then return Left
21Rational agentsFor each possible percept sequence, a rational agent should select an action that is expected to maximize its performance measure, given the evidence provided by the percept sequence and the agent’s built-in knowledgePerformance measure An objective criterion for success of an agent's behavior
22What does rationality mean? Rationality is not omnisciencePercepts may not supply all the relevant informationConsequences of actions may be unpredictableAn agent is autonomous if its behavior is determined by its own experience (with ability to learn and adapt)
24Agent: Part-sorting robot Performance measurePercentage of parts in correct binsEnvironmentConveyor belt with parts, binsActuatorsRobotic armSensorsCamera, joint angle sensors
25Agent: Spam filter Performance measure Environment Actuators Sensors Minimizing false positives, false negativesEnvironmentA user’s accountActuatorsMark as spam, delete, etc.SensorsIncoming messages, other information about user’s account
26Environment typesFully observable (vs. partially observable): The agent's sensors give it access to the complete state of the environment at each point in timeDeterministic (vs. stochastic):The next state of the environment is completely determined by the current state and the agent’s actionEpisodic (vs. sequential):The agent's experience is divided into atomic “episodes,” and the choice of action in each episode depends only on the episode itself
27Environment types Static (vs. dynamic): Discrete (vs. continuous): The environment is unchanged while an agent is deliberatingSemidynamic: the environment does not change with the passage of time, but the agent's performance score doesDiscrete (vs. continuous):The environment provides a fixed number of distinct percepts, actions, and environment statesTime can also evolve in a discrete or continuous fashion
28Environment typesSingle agent (vs. multi-agent): An agent operating by itself in an environment Known (vs. unknown): The agent knows the rules of the environment
29Examples Fully observable Deterministic Episodic Static Discrete Chess w Clock without clock TaxiFully observableDeterministicEpisodicStaticDiscreteSingle AgentYesYesNoStrategicNoStrategicNoNoNoSemiYesNoYesYesNoNoNoNo
31Simple reflex agentSelect action on the basis of current percept, ignoring all past percepts
32Model-based reflex agent Maintains internal state that keeps track of aspects of the environment that cannot be currently observed
33Goal-based agentThe agent uses goal information to select between possible actions in the current state
34Utility-based agentThe agent uses a utility function to evaluate the desirability of states that could result from each possible action
35A learning agentto build learning machines and then to teach them.
36learning element, which is responsible for making improvements, A learning Agentlearning element, which is responsible for making improvements,performance element, which is responsible for selecting external actions.The learning element uses feedback from the critic on how the agent is doing and determines how the performance element should be modified to do better in the future.problem generator: responsible for suggesting actions that will lead to new and informative experiences