Set of subjects for the written exam at “Artificial Intelligence”.

The list is complete. References are:

[1] Russel and Norvig book from the lectures bibliography

[2] Russel lecture notes

[3] Giarratano-Riley book from the bibliography

[4] my lecture notes.

[5] wikipedia

1) AI definitions.

2) The Turing test.

3) What are the cognitive sciences and the relation between them and AI.

4) What are the roots of AI.

5) Name at least 3 important facts from the history of AI.

6) The General Problem Solver.

7) The Mycin system

8) The Eliza system.

9) Name at least 3 programming languages and/or frameworks specific for AI. (Note: general purpose programming languages, such as Java are excluded).

10) Name at least 5 subfields of AI.

11) Describe the Breadth-First search algorithm. Example.

12) Describe the Depth-First search algorithm. Example.

13) Describe the A* search algorithm. Example.

14) Describe the IDA* search algorithm. Example.

15) Describe the SMA* search algorithm. Example.

16) Describe the RBFS search algorithm. Example.

17) Describe the Hill-climbing search algorithm. Example.

18) Describe the Genetic algorithm principle.

19) Describe the MIN-MAX algorithm. Example.

20) Describe the Alpha-Beta pruning algorithm. Example.

21) Describe the Constraint Satisfying Problem (CSP).

22) Describe the Backtracking search algorithm. Example.

23) Describe the Distributed Constraint Satisfying Problem (DCSP).

24) Describe the Asynchronous BackTracking (ABT) Family of algorithms.

25) Knowledge Base (KB). Definition. Describe the KB agent’s actions.

26) Expert system. Definition and working principles.

27) Knowledge representation using the first order predicate logic. Constants, predicates, functions, variables, connectives, quantifiers.

28) The modus ponens inference rule. Example.

29) The modus tolens inference rule. Example.

30) Forward chaining algorithm. General description.

31) Backward chaining algorithm. General description.

32) Knowledge representation using production rules. Example.

33) Knowledge representation using semantic networks. Example.

34) Knowledge representation using Frames. Example.

35) Knowledge representation using decision trees. Example.

36) Knowledge representation using AND-OR trees. Example.

37) The OAV model for representing semantic nets. Example.

38) CLIPS/JESS. Structured facts (defined with deftemplate). Example

39) CLIPS/JESS. Unstructured facts. Example

40) CLIPS/JESS. Assert and retract. Examples.

41) CLIPS/JESS. Constraints. Examples

42) CLIPS/JESS. Relational expressions. Examples.

43) CLIPS/JESS. Reading data from the keyboard.

44) CLIPS/Jess. Loop structures and/or techniques.

45) Planning. The language of planning problems (from section 11.1 of [1]).

46) Planning. Forward state-space search

47) Planning. Backward state-space search.

48) Bayes' Rule and Its Use (from section 13.6 of [1]).

49) The Semantics of Bayesian Networks (from section 14.2 of [1]).

50) Inference in Temporal Models (from section 15.2 of [1]).

51) Hidden Markov Models (from section 15.3 of [1]).

52) Machine learning. Forms of learning (from section 18.1 of [1]).

53) Machine learning. Inductive learning (from section 18.2 of [1])

54) Machine learning. Learning decision trees algorithm (from section 18.3 of [1])

55) Statistical Learning Methods. Naïve Bayes models (from section 20.2 of [1])

56) Single layer feed-forward neural networks (perceptrons) (from section 20.5 of [1])

57) Multilayer feed-forward neural networks (from section 20.5 of [1]).

The list of theoretical subjects (1-57) is final. Late additions to the list are marked with red “ink”.

In the list will be included also sample problems to solve in CLIPS/JESS. The problems you will receive at the exam will be similar to those presented here. It may be that in the list are also some problems that are not CLIPS/JESS related, but are about the first part of the lecture, the search algorithms, or the knowledge representation part.

Some CLIPS problems might resemble the ones from the second assignment.

Some search algorithms problems might resemble the problems from the first assignment and/or those from the corresponding chapters from the Russel and Norvig book, but for the problems from the first assignment, they most probably will only require to draw the state space (tree) and label (with a number) the nodes in the order that the algorithm will process them.

Knowledge representation problems may sond like:

1) Draw a {decision tree | AND-OR tree | semantic network} for for the following problem …

2) Write a CLIPS program for the following {decision tree | AND-OR tree | semantic network} …

This version was saved at 23:55, 17.01.2014