What Is On he inal Review Everything that has important! written next to it on the slides Everything that I said was important ECE457 Applied Artificial Intelligence all 27 ecture #14 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 2 What Might Be On he inal Anything in the slides Except What Is Not On he inal Anything in the required readings in the textbook What Is Not On he inal Examples of real applications Dune II raveling-wave tube IBM Deep Blue Pathfinder network Weighted Naïve Bayes Classifier Helicopter flight control Neural network pixel classifier uzzy robot navigation WordNet Additional material on website Writing/debugging code ECE457 Applied Artificial Intelligence R. Khoury (27) Page 3 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 4 1
Practice Material Examples and exercises in slides Problems at the end of each chapter in the textbook Material Allowed at the Exam Pen or pencil, eraser, calculator Not allowed: Books, notes Phones, blackberries, laptops, PDAs, ipods, iphones, ianything, computers built into glasses like in Mission Impossible, or anything else electronic alking to other students, writing notes, sign language, smoke signals, semaphores Cheating in general ECE457 Applied Artificial Intelligence R. Khoury (27) Page 5 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 6 ecture 1: Introduction to AI ypes of agents Properties of the environment ecture 1: Introduction to AI Define the properties of the environment for these problems: Robot soccer Internet shopping (without ebay-style bidding) Autonomous Mars rover heorem-solving assistant to a mathematician ECE457 Applied Artificial Intelligence R. Khoury (27) Page 7 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 8 2
ecture 2: Uninformed earch Well-defined problem Properties of search algorithms Uninformed search Breath-first search Uniform-cost search Depth-first search Depth-limited search Iterative deepening search Repeated states ECE457 Applied Artificial Intelligence R. Khoury (27) Page 9 ecture 2: Uninformed earch You have a search tree with a branching factor of b and a maximum depth of m. he depth of the shallowest goal node is d. You are considering searching the tree using either a depth-first search agent or a breathfirst search agent. Which one will have the best space complexity? Explain. ECE457 Applied Artificial Intelligence R. Khoury (27) Page 1 ecture 2: Uninformed earch You have a search tree with a branching factor of b and a maximum depth of m. he depth of the shallowest goal node is d. You are considering searching the tree using either a depth-first search agent or a breathfirst search agent. Which one will have the best time complexity? Explain. ecture 2: Uninformed earch A 3-foot-tall monkey is in a room where some bananas are suspended from the 8-foot-high ceiling. He would like to get the bananas as quickly as possible. he room contains two stackable, movable climbable 3-foot-high crates. Write this situation as a well-defined problem. ECE457 Applied Artificial Intelligence R. Khoury (27) Page 11 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 12 3
ecture 2: Uninformed earch Initial state Action Goal test Cost ecture 3: Informed earch Informed search Greedy best-first search A* search Heuristic functions Iterative improvement Hill Climbing imulated Annealing ECE457 Applied Artificial Intelligence R. Khoury (27) Page 13 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 14 ecture 3: Informed earch Given the following tree, find the optimal path to the goal G using A* search. he value of the heuristic h is specified for each node. he costs of the edges are specified on the tree. Assume that children of a node are placed into the list in a left-to-right order, and that nodes of equal priority are extracted (for expansion) from the list in IO order. Write a number inside the node indicating the order in which the nodes are expanded from the start node, i.e. 1, 2,. or each node generated, write the total cost f in the appropriate location on the graph. ecture 3: Informed earch ind the optimal path to the goal G using A* search, specifying the order in which nodes are expanded and the f-value of all generated nodes. ECE457 Applied Artificial Intelligence R. Khoury (27) Page 15 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 16 4
ecture 4: CP ecture 4: Constraint atisfaction Problems Constraints Defining a CP CP search Backtracking search Conflict-directed backjumping Heuristics orward checking AC-3 algorithm C A E Using the most-constrained-variable CP heuristic, colour the adjacent map using the colours Blue, Red and Green. how your reasoning at each step of the algorithm. Proceed in the following manner: After assigning a colour to a region, and before choosing the next region to colour, apply the forward checking algorithm and show its results. hen choose the next region to colour using the B most-constrained-variable heuristic, etc. At each step, show the domains of each region and justify the choice of the next region to colour. ECE457 Applied Artificial Intelligence R. Khoury (27) Page 17 ECE457 DApplied Artificial Intelligence R. Khoury (27) Page 18 A ecture 4: CP B Variables marked * have been assigned A* = {Green} B* = {Red} C = {Red, Blue, Green} D = {Red, Blue, Green} E = {Red, Blue, Green} = {Red, Blue, Green} ecture 5: Game Playing Payoff functions Minimax algorithm Alpha-Beta pruning Non-quiescent positions & horizon effect Expectiminimax C E ECE457 DApplied Artificial Intelligence R. Khoury (27) Page 19 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 2 5
ecture 5: Game Playing Consider the following game tree. he payoff value of each leaf is written under that node. Apply the Minimax algorithm to obtain the value of each non-leaf node. Apply Alpha-Beta Pruning to the game tree. ind which nodes will be pruned. or each one, identify and explain the value of alpha and beta to show why it is pruned. ecture 5: Game Playing MAX MIN A B H MAX C I D E G H J K M N 4 8 9-1 -2 2-8 5 A B C H I ECE457 Applied Artificial Intelligence R. Khoury (27) Page 21 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 22 ecture 6: ogical Agents anguage, syntax, semantics Propositional logic Propositional symbols and logical connectives Inference with truth tables Inference with Resolution Conversion to CN Inference with Modus Ponens Horn clauses orward chaining Backward chaining ECE457 Applied Artificial Intelligence R. Khoury (27) Page 23 ecture 7: irst-order ogic irst-order ogic Constants, predicates, functions Universal and existential quantifiers Converting English sentences Inference with propositionalization Inference with Generalized Modus Ponens Unification algorithm Inference with Resolution Conversion to CN ECE457 Applied Artificial Intelligence R. Khoury (27) Page 24 6
ecture 7: irst-order ogic Represent the following sentences in O using: ake(s,c,t), Pass(s,c,t), core(s,c,t), tudent(s), rench, Greek, pring21 ome students took rench in spring 21 Every student who takes rench passes it ecture 7: irst-order ogic Convert this O sentences to Conjunctive Normal orm. how all steps of the conversion. x [ y (y) G(x,y)] y G(y,x) Only one student took Greek in pring 21 he best score in Greek is always higher than the best score in rench ECE457 Applied Artificial Intelligence R. Khoury (27) Page 25 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 26 ecture 7: irst-order ogic ind the most general unifier, if it exists. p = (A,B,B) q = (x,y,z) p = (y,g(a,b)) q = (G(x,x),y) p = (G(y),y) q = (G(x),A) ecture 7: irst-order ogic Given the following KB: aster(x,y) aster(y,z) aster(x,z) Pig(x) lug(y) aster(x,y) Buffalo(x) Pig(y) aster(x,y) lug(limm) Pig(Pat) Buffalo(Bill) Is Bill faster than limm, using forward chaining p = (G(y),y) q = (x,x) ECE457 Applied Artificial Intelligence R. Khoury (27) Page 27 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 28 7
ecture 7: irst-order ogic Given the following KB: limy(x) Creepy(x) lug(x) Pig(x) lug(y) aster(x,y) limy(limm) Creepy(limm) Pig(Pat) Is Pat faster than limm, using backward chaining Given the following KB: Person(Marcus) Pompeian(Marcus) Ruler(Caesar) Assasinate(Marcus, Caesar) ecture 7: irst-order ogic Pompeian(x1) Roman(x1) Roman(x2) oyal(x2,caesar) Hate(x2, Caesar) Person(x3) Ruler (x4) Assasinate(x3, x4) oyal(x3,x4) Does Marcus hate Caesar, using resolution ECE457 Applied Artificial Intelligence R. Khoury (27) Page 29 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 3 ecture 8: Uncertainty Marginalization Bayes heorem Chain rule Independence and conditional independence Naïve Bayes Classifier ecture 8: Uncertainty You tested positive for a disease. he test s results are accurate 99% of the time. However, the disease only strikes 1 out of 1 people. What s the probability that you have the disease? ECE457 Applied Artificial Intelligence R. Khoury (27) Page 31 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 32 8
ecture 8: Uncertainty Given the following police data, create a Naïve Bayes Classifier for stolen cars, and compute the probability that a domestic red UV is stolen. C Red Red Red Yellow Yellow Yellow Yellow Yellow Red ports ports ports ports ports UV UV UV UV O Domestic Domestic Domestic Domestic Imported Imported Imported Domestic Imported tolen Not tolen Not tolen Not tolen Not Not ecture 9: Probabilistic Reasoning Bayesian Network Connections and D-eparation Inference Red ports Imported tolen ECE457 Applied Artificial Intelligence R. Khoury (27) Page 33 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 34 ecture 9: Probabilistic Reasoning Consider this Bayesian network. Write the factored expression for the joint probability distribution P(A, B, C, D, E, ) which is represented by this network. Which variables are independent (dseparate) of C if: B is known. A is known. D and E are both know. ecture 9: Probabilistic Reasoning Given the following values, what is the posterior probability of given that B is true? P(D B) =.8 P(D B) =.4 P( D) =.75 P( D) =.6 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 35 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 36 9
ecture 1: Decision Making ecture 1: Decision Making Maximum Expected Utility Utility Expected utility Decision network Optimal policy Computing the optimal policy Value of information Wear Protection (Pr) W Which Way (W) P(A).6 Accident (A) U Pr W A U.4.7.3.6.1 1.9.3 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 37 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 38 ecture 1: Decision Making l i P(G).9 P( G).1 P(N) 1 1 Utility cost of inspection = -5 l b i U -3 ecture 11: Introduction to earning or all learning algorithms raining data Objective of learning emon P() =.5.2.8 Report Inspect Buy U -3 1-6 -35-35 95-65 Evaluation General algorithm Precision and recall Overfitting and n-fold cross-validation K-Means Q-earning Exploration function ECE457 Applied Artificial Intelligence R. Khoury (27) Page 43 39 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 55 4 1
ecture 12: Introduction to oft Computing Artificial neural networks Artificial neuron Perceptron network uzzy logic uzzy sets, fuzzy membership functions, membership degree uzzy rules Genetic algorithms Individuals Operators: crossover, mutation, selection earch algorithm ecture 13: Introduction to Ontologies Objects, Categories, Relations, Attributes Inheritance Problems ECE457 Applied Artificial Intelligence R. Khoury (27) Page 56 41 ECE457 Applied Artificial Intelligence R. Khoury (27) Page 57 42 11