CS 404 Artificial Intelligence Review

1
CS 404 Artificial IntelligenceReview
2
Chp. 3,4 Searching
Continental Airlines Route Map
USAir America West Route Map
3
Chp. 3, 4 Searches

• Blind Searches
• Breadth-first search
• Depth-first search
• Iterative-depening search
• Uniform-cost search
• Bidirectional search
• Informed Searches
• Greedy search
• A search

• Heuristic issues
• What makes a good heuristic (admissable,
  consistent,)
• How to devise heuristics
• Effective branching factor
• Graph search versus Tree search

4
Chp. 3, 4 Searches

• Local Searches
• Hill-Climbing
• Hill-Climbing variations
• stochastic hill-climbing
• first choice hill-climbing
• random restart
• Simulated Annealing
• Local beam search
• Genetic Algorithms
• When to prefer to exhaustive search techniques

5
Chp 5 Constraint Satisfaction Problems

• Blind or Informed Searches may be applicable
• Heuristics are more general
• most constrained variable
• least constraining value
• Success rate is high

6
Chp 6 Game Playing

• Deterministic games
• Minimax
• Minimax with Alpha-Beta pruning
• Non-deterministic games
• Expectimax
• Value of look-ahead
• State-of-the-art

7
Chp 7 Propositional Logic

• Concepts
• entailment
• models, worlds
• Syntax and semantics
• Logical Equivalence
• Validity and Satisfiability
• Inference
• truth table method
• application of inference rules
• Modus Ponens,
• Restricted forms
• CNF
• Resolution

8
Chp 8,9 First Order Logic

• Added to Propositional Logic
• Existential and Universial Quantifiers
• Unification
• Conversion to Normal Forms
• Conjunctive Normal Form (CNF)
• Implicative Normal Form
• Generalized Modus Ponens
• Forward, Backward chaining
• Resolution

9
Chp 13 Uncertainty

• Definition of
• Prior probabilities
• Joint probabilities
• Conditional probabilities
• Bayes Rule
• Summing over hidden variables
• Independence

10
Chp 14 Bayesian Belief Networks

• Constructing
• using conditional indepences
• Computing probabilities
• updating beliefs
• Stochastic sampling methods
• How can you estimate the joint probability of an
  event using stochastic sampling

11
Chp 16 Rational Decisions

• Concept of a lottery
• Certainty equivalent of a lottery
• Utility of money
• Making rational decisions
• Showing your alternatives as a tree
• Taking best actions when necessary

12
Chp 18 Learning

• Goals of learning
• classification
• regression
• Performance criteria
• error on test/unseen data -gt generalization
  performance
• error measures
• sum of squared error
• mean squared error
• Complexity of models
• how it affects generalization performance
• Occams razor
• Decision Trees
• capabilities disjuncction of conjunctions
• how to select attributes for each node concept
  of entropy
• what is overfitting
• what is pruning, how to use a pruned tree

13
What we have covered

• Part I Artificial Intelligence      1
  Introduction      2 Intelligent Agents
• Part II Problem Solving      3 Solving Problems
  by Searching      4 Informed Search and
  Exploration      5 Constraint Satisfaction
  Problems     6 Adversarial Search
• Part III Knowledge and Reasoning      7 Logical
  Agents      8 First-Order Logic      9
  Inference in First-Order Logic     10 Knowledge
  Representation
• Part IV Planning     11 Planning     12
  Planning and Acting in the Real World
• Part V Uncertain Knowledge and Reasoning     13
  Uncertainty     14 Probabilistic Reasoning
      15 Probabilistic Reasoning Over Time     16
  Making Simple Decisions     17 Making Complex
  Decisions
• Part VI Learning     18 Learning from
  Observations    19 Knowledge in Learning     20
  Statistical Learning Methods (overview only)
•     21 Reinforcement Learning

• Part VII Communicating, Perceiving, and Acting
      22 Communication     23 Probabilistic
  Language Processing     24 Perception     25
  Robotics
• Part VIII Conclusions     26 Philosophical
  Foundations     27 AI Present and Future
• (covered somewhat in the 1st lecture)

14
Evaluations

• Please fill out your course evaluations
• They are always taken into account
• They are anonymous
• You can improve the course
• Try to
• enter verbal remarks
• more useful than trying to judge atmosphere of
  the learning environment
• if you want better homeworks/exams, try to
  indicate what you mean
• be balanced in your comments
• e.g. if we should cover another topic or go
  slower, tell us what we should drop
• what is going right and should not change?