Artificial Intelligence COMP-424 - McGill University School Of Computer .
Artificial IntelligenceCOMP-424Lecture notesby Alexandre TombergProf. Joelle PineauMcGill UniversityWinter 2009Lecture notes Page 1
Table of ContentsDecember-03-0812:16 PMI. History of AIII. Search1. Uninformed Search Methods2. Informed Search3. Search for Optimization Problems4. Game Playing5. Constraint SatisfactionIII. Logic1. Knowledge Representation: Logic2. First Order Logic3. Planning4. Spatial PlanningIV. Probability1. Reasoning under Uncertainty2. Bayesian NetworksV. Machine Learning1. Machine Learning: Parameter Estimation2. Learning with Missing Values3. Supervised Learning4. Neural Nets5. Decision TreesVI. Decision Theory1. Utility Theory2. Markov Decision Processes (MDPs)3. Reinforcement LearningLecture notes Page 2
History of AIJanuary-06-0910:03 AMLecture notes Page 3
Uninformed Search MethodsJanuary-08-0910:06 AMLecture notes Page 4
Lecture notes Page 5
Generic Search Algorithm:Algorithm 1: BFSLecture notes Page 6
Algorithm 2: DFSAlgorithm 3: Depth limited searchAlgorithm 4: Iterative DeepeningLecture notes Page 7
Informed SearchJanuary-13-0910:02 AMLecture notes Page 8
AlgorithmsJanuary-13-0910:34 AMAlgorithm #1: Best-First SearchAlgorithm #2: Heuristic SearchLecture notes Page 9
Algorithm # 3: A* searchLecture notes Page 10
Lecture notes Page 11
Search for Optimization ProblemsJanuary-15-0910:05 AMLecture notes Page 12
Iterative Improvement AlgorithmsJanuary-15-0910:05 AMAlgorithm #1: Hill ClimbingAlgorithm #2: Simulated AnnealingLecture notes Page 13
Lecture notes Page 14
Genetic AlgorithmsJanuary-15-0911:06 AMLecture notes Page 15
Game PlayingJanuary-20-0910:03 AMLecture notes Page 16
Minimax SearchJanuary-20-0910:07 AMLecture notes Page 17
α-β PruningJanuary-20-0910:44 AMLecture notes Page 18
Constraint SatisfactionJanuary-22-0910:10 AMLecture notes Page 19
Lecture notes Page 20
Lecture notes Page 21
Knowledge Representation: LogicJanuary-27-0910:10 AMLecture notes Page 22
Lecture notes Page 23
Lecture notes Page 24
Lecture notes Page 25
Lecture notes Page 26
Lecture notes Page 27
First Order LogicFebruary-18-097:50 PMLecture notes Page 28
Lecture notes Page 29
Lecture notes Page 30
Lecture notes Page 31
Lecture notes Page 32
Lecture notes Page 33
Lecture notes Page 34
PlanningFebruary-03-0910:11 AMLecture notes Page 35
Lecture notes Page 36
Lecture notes Page 37
Lecture notes Page 38
Partial Order Planning AlgorithmFebruary-18-098:55 PMLecture notes Page 39
Least CommitmentAnalysisLecture notes Page 40
Spatial PlanningFebruary-03-0910:32 AMLecture notes Page 41
Lecture notes Page 42
Lecture notes Page 43
Lecture notes Page 44
Reasoning under UncertaintyFebruary-18-099:13 PMIf we know probabilities, what actions should we choose?Lecture notes Page 45
Lecture notes Page 46
Lecture notes Page 47
Lecture notes Page 48
Lecture notes Page 49
Lecture notes Page 50
Bayesian NetworksMarch-19-093:26 PMLecture notes Page 51
Lecture notes Page 52
Machine Learning: Parameter EstimationMarch-03-0910:09 AMLecture notes Page 53
Statistical Parameter FittingMarch-03-0910:34 AMLecture notes Page 54
Maximum Likelihood Estimate (MLE)March-03-0910:53 AMLecture notes Page 55
Lecture notes Page 56
Learning with Missing ValuesMarch-10-0910:14 AMLecture notes Page 57
Basic EM algorithm:Start with an initial parameter settingRepeat:Expectation Step: Complete the data by assigning values to missingitems.Maximization Step: Compute the maximum log-likelihood and newparameters on the complete data.Lecture notes Page 58
Lecture notes Page 59
Soft EM for a general Bayes net:Lecture notes Page 60
Machine Learning: ClusteringMarch-19-094:21 PMLecture notes Page 61
Lecture notes Page 62
Supervised LearningMarch-10-0910:55 AMLecture notes Page 63
Lecture notes Page 64
Lecture notes Page 65
OverfittingApril-14-098:35 PMLecture notes Page 66
Lecture notes Page 67
Finding Parameters in GeneralApril-14-099:05 PMGradient Descent:Given w0, for i 0, 1, 2, . do:Repeat until necessary.Lecture notes Page 68
Batch vs. Online OptimizationApril-14-099:38 PMLecture notes Page 69
What we should know:Lecture notes Page 70
Neural NetsMarch-19-094:48 PMLecture notes Page 71
Lecture notes Page 72
Lecture notes Page 73
Lecture notes Page 74
Feed Forward Neural NetworksApril-15-0910:48 AMForward pass:for layer k 1 . K do:Compute the output of all units in layer kCopy this output as the input to the next layerLecture notes Page 75
Lecture notes Page 76
Backpropagation algorithm:1. Forward pass: compute the output of the network going from inputlayer to output layer.2. Backward pass: compute the gradient of the error for every weightinside the network going from output layer towards the input layer.3. Update: update the weights using the standard rule:Lecture notes Page 77
Lecture notes Page 78
Overfitting in Neural NetApril-15-0912:56 PMLecture notes Page 79
Decision TreesApril-15-091:04 PMLecture notes Page 80
Lecture notes Page 81
Lecture notes Page 82
Lecture notes Page 83
Lecture notes Page 84
Utility TheoryApril-15-091:54 PMLecture notes Page 85
Utility Models:Lecture notes Page 86
Maximizing Expected Utility (MEU) PrincipleApril-15-092:21 PMLecture notes Page 87
Lecture notes Page 88
What we should know:Lecture notes Page 89
Markov Decision Processes (MDPs)April-15-092:50 PMLecture notes Page 90
Lecture notes Page 91
PoliciesApril-15-092:50 PMLecture notes Page 92
Lecture notes Page 93
Iterative Policy Evaluation Algorithm:1. Start with some initial guess2. During iteration k update the function for all states as follows:Lecture notes Page 94
Searching for a Good PolicyApril-15-094:47 PMLecture notes Page 95
Policy Iteration Algorithm:Start with an initial policyRepeat untilComputeusing policy evaluation algorithmComputeusing greedy policy update rule onLecture notes Page 96
Value Iteration Algorithm:Start with an initial valueRepeat untilUpdate the value function estimate using:Lecture notes Page 97
Lecture notes Page 98
Lecture notes Page 99
Reinforcement LearningApril-15-095:38 PMLecture notes Page 100
Lecture notes Page 101
TD (order 0) Learning Algorithm:1. Initialize the value function:2. Repeat until feeling sick of it:a. Pick a start stateb. Repeat for every time step ti. Choose an action a based on current policy π and current state sii. Take action a, observe reward r and new state s'iii. Compute TD error: δ r γ V(s') - V(s)iv. Update the value function: V(s) V(s) αs δv. Update current state: s s'vi. If s' is a terminal state, GoTo 2.Lecture notes Page 102
Reinforcement Learning for ControlApril-15-096:35 PMLecture notes Page 103
Lecture notes Page 104
Artificial Intelligence COMP-424 Lecture notes by Alexandre Tomberg Prof. Joelle Pineau McGill University Winter 2009 Lecture notes Page 1 . I. History of AI 1. Uninformed Search Methods . Lecture notes Page 58 . Lecture notes Page 59 . Soft EM for a general Bayes net: Lecture notes Page 60 . Machine Learning: Clustering March-19-09
COMP-424: Artificial intelligence 8 Joelle Pineau Example: Robot path planning State space: robot's position Operators: go-north, go-south, go-east, go-west Goal: target location Path cost: path length COMP-424, Lecture 2 - January 9, 2013 5 Defining a Search Problem State spaceS: all possible configurations of the .
Artificial Intelligence -a brief introduction Project Management and Artificial Intelligence -Beyond human imagination! November 2018 7 Artificial Intelligence Applications Artificial Intelligence is the ability of a system to perform tasks through intelligent deduction, when provided with an abstract set of information.
Bearing, invented by McGill over 60 years ago CAMROL is the trademark for the line of roller bearing cam followers manufactured by McGill and de-signed for use as cam followers and track rollers. The CAMROL Cam Follower Bearing was originally in-vented by McGill over 60 years ago. Since then, McGill has maintained its leading position through the
Song of St. Patrick – Haugen – G Comp II # 685 Taste and See – Haugen – G Comp II # 34 Taste and See – Moore – G Comp II # 827 The Love of The Lord – Joncas – G Comp II # 680 The Servant Song – Richard Gillard– – G Comp II # 661 We Have Been Told – Haas – G Comp II # 69
2016-17 HERI Faculty Survey Institutional Profile Report Full-time Undergraduate Faculty Total Men Women CIRP Construct Note: Significance * p .05, ** p .01, *** p .001 Page 1 of 76 1A. American University of Beirut Your Inst Comp 1 Comp 2 Your Inst Comp 1 Comp 2 Your Inst Comp 1 Comp 2
and artificial intelligence expert, joined Ernst & Young as the person in charge of its global innovative artificial intelligence team. In recent years, many countries have been competing to carry out research and application of artificial intelli-gence, and the call for he use of artificial
BCS Foundation Certificate in Artificial Intelligence V1.1 Oct 2020 Syllabus Learning Objectives 1. Ethical and Sustainable Human and Artificial Intelligence (20%) Candidates will be able to: 1.1. Recall the general definition of Human and Artificial Intelligence (AI). 1.1.1. Describe the concept of intelligent agents. 1.1.2. Describe a modern .
Akuntansi manajemen mempunyai peranan besar dalam perusahaan, yaitu membantu pihak pihak internal (direktur utama dan masing masing tingkatan manajer dalam setiap unit/departemen) dalam pengambilan keputusan. Oleh karena itu, akuntansi manajemen yang akan kita pelajari dalam buku ini akan membahas hal hal sebagai berikut: 1. Konsep dan fungsi biaya Pihak manajemen dapat memahami berbagai .
