1. Topics in Artificial Intelligence
prof. dr hab. inż. Joanna Józefowska,

2. Curriculum
Introduction – overview of research topics in artificial intelligence
Knowledge representation
Space search as a general inference model
Reasoning under uncertainty
dr hab. inż. Joanna Józefowska, prof. PP

3. References dr hab. inż. Joanna Józefowska, prof. PP
Bolc L., Borodziewicz W., Wójcik M., Podstawy przetwarzania informacji niepewnej i niepełnej, PWN, Warszawa, 1991.
Bolc L., Zaremba J., Wprowadzenie do uczenia się maszyn, Akademicka Oficyna Wydawnicza RM, Warszawa, 1992.
Bolc L., J. Cytowski, Metody przeszukiwania heurystycznego, PWN, t1 1989, t
Charniak E., Mc Dermot D., Introduction to Artificial Intelligence, Addison Wesley, 1985.
Churchland P.M., P. Smith-Churchland, Czy maszyna może myśleć?, Świat Nauki, lipiec 1991.
Greenfield S., Tajemnice mózgu, Świat Książki, Warszawa, 1998.
Guida G., C. Tasso, Design and Development of Knowledge-Based Systems, John Wiley 1994.
Harel D., Rzecz o istocie informatyki, wyd. 2, WNT Warszawa, 2000.
Lugger G., Stubblefield W.A., Artificial Intelligence and the Design of Expert Systems, The Benjamin/Cummings Publ. Comp. Inc., 1989.
Mulawka J., Systemy ekspertowe, Warszawa, WNT, 1996
Neural Networks and Soft Computing, L. Rutkowski, R. Tadeusiewicz (eds.), Polish Neural Network Society, Częstochowa, 2000.
Niederliński A., Regułowe systemy ekspertowe, Wydawnictwo Pracowni Komputerowej Jacka Skalmierskiego, Gliwice 2000.
Puppe F., Systematic Introduction to Expert Systems, Springer Verlag 1993.
Rich E., Artificial Intelligence, McGraw Hill, 1983.
Rich E., K. Knight, Artificial intelligence, McGraw Hill, New York, 1991.
Russell S. J., Norvig P., Artificial Intelligence. A modern approach, Prentice Hall, Inc
Scarle J.R., Czy intelekt mózgu jest programem komputerowym?, Świat Nauki, lipiec 1991.
Sieci Neuronowe, W. Duch, J. Korbicz, L.Rutkowski, R. Tadeusiewicz, Biocybernetyka i Inżynieria Medyczna 2000, t. 6, Akademicka Oficyna Wydawnicza EXIT, Warszawa 2000.
Tadeusiewicz R., Elementarne wprowadzenie do techniki sieci neuronowych z przykładowymi programami, Akademicka Oficyna Wydawnicza PLJ, Warszawa 1998.
dr hab. inż. Joanna Józefowska, prof. PP

4. Artificial Intelligence myths and reality
dr hab. inż. Joanna Józefowska, prof. PP

5. What is human intelligence?
Is it a single feature or a set of skills?
Can one learn it?
What is learning?
What is creativity?
What is intuition?
What is consciousness?
Can we build an intelligent machine?
How to check if a machine is intelligent?
dr hab. inż. Joanna Józefowska, prof. PP

6. Intelligence has been defined by prominent researchers in the field as :
Binet and Simon (1905): the ability to judge well, to understand well, to reason well.
Terman (1916): the capacity to form concepts and to grasp their significance.
Wechsler (1939): the aggregate or global capacity of the individual to act purposefully, to think rationally, and to deal effectively with the environment.
Gardner (1986): the ability or skill to solve problems or to fashion products which are valued within one or more cultural settings.
dr hab. inż. Joanna Józefowska, prof. PP

7. Linguistic intelligence
reading
writing
speaking
understanding
creativity
Shall I compare thee to a summer’s day?
Thou art. More lovely and more temperate:
Rough winds do shake the darlings buds of May,
And summer’s lease hath all too short a date;
W. Shakespeare
Linguistic intelligence is the ability to think in words and to use language to express and appreciate complex meanings.  Linguistic intelligence allows us to understand the order and meaning of words and to apply meta-linguistic skills to reflect on our use of language.  Linguistic intelligence is the most widely shared human competence and is evident in poets, novelists, journalists, and effective public speakers.  Young adults with this kind of intelligence enjoy writing, reading, telling stories or doing crossword puzzles.
dr hab. inż. Joanna Józefowska, prof. PP

8. Personal intelligence
Ma dwie odmiany:
interpersonal
– „people smart”
intrapersonal
–„self smart”
Interpersonal intelligence is the ability to understand and interact effectively with others.  It involves effective verbal and nonverbal communication, the ability to note distinctions among others, sensitivity to the moods and temperaments of others, and the ability to entertain multiple perspectives.  Teachers, social workers, actors, and politicians all exhibit interpersonal intelligence.  Young adults with this kind of intelligence are leaders among their peers, are good at communicating, and seem to understand others’ feelings and motives.
Intra-personal intelligence is the capacity to understand oneself and one’s thoughts and feelings, and to use such knowledge in planning and directioning one’s life.  Intra-personal intelligence involves not only an appreciation of the self, but also of the human condition.  It is evident in psychologist, spiritual leaders, and philosophers.  These young adults may be shy.  They are very aware of their own feelings and are self-motivated.
dr hab. inż. Joanna Józefowska, prof. PP

9. Logical-mathematical intelligence „number/reasoning smart”
abstract, symbolic thought
sequential reasoning skills
inductive and deductive thinking patterns
E=mc2
Logical-mathematical intelligence is the ability to calculate, quantify, consider propositions and hypotheses, and carry out complete mathematical operations.  It enables us to perceive relationships and connections and to use abstract, symbolic thought; sequential reasoning skills; and inductive and deductive thinking patterns.  Logical intelligence is usually well developed in mathematicians, scientists, and detectives.  Young adults with lots of logical intelligence are interested in patterns, categories, and relationships.  They are drawn to arithmetic problems, strategy games and experiments.
Ampere i Gauss byli genialnymi mnemotechnikami w dzieciństwie, ale
potem te zdolności zaczęły zanikać
dr hab. inż. Joanna Józefowska, prof. PP

10. Kinesthetic intelligence – „Body smart”
manipulate objects and use a variety of physical skills
Bodily kinesthetic intelligence is the capacity to manipulate objects and use a variety of physical skills.  This intelligence also involves a sense of timing and the perfection of skills through mind–body union. 
dr hab. inż. Joanna Józefowska, prof. PP

11. Słuchamy fragmentu IX symfonii Ludwiga van Beethovena
Musical intelligence
Musical intelligence is the capacity to discern pitch, rhythm, timbre, and tone. 
recognize
create
reproduce
reflect on music
Musical intelligence is the capacity to discern pitch, rhythm, timbre, and tone.  This intelligence enables us to recognize, create, reproduce, and reflect on music, as demonstrated by composers, conductors, musicians, vocalist, and sensitive listeners.  Interestingly, there is often an affective connection between music and the emotions; and mathematical and musical intelligences may share common thinking processes.  Young adults with this kind of intelligence are usually singing or drumming to themselves.  They are usually quite aware of sounds others may miss.
Słuchamy fragmentu IX symfonii Ludwiga van Beethovena
dr hab. inż. Joanna Józefowska, prof. PP

12. Spatial intelligence mental imagery spatial reasoning
Spatial intelligence is the ability to think in three dimensions. 
mental imagery
spatial reasoning
image manipulation
graphic and artistic skills
an active imagination
Spatial intelligence is the ability to think in three dimensions.  Core capacities include mental imagery, spatial reasoning, image manipulation, graphic and artistic skills, and an active imagination.  Sailors, pilots, sculptors, painters, and architects all exhibit spatial intelligence.  Young adults with this kind of intelligence may be fascinated with mazes or jigsaw puzzles, or spend free time drawing or daydreaming.
dr hab. inż. Joanna Józefowska, prof. PP

13. The IQ Test and 7 types of intelligence by Gardner
IQ = (Mental Age) / (Chronological Age) x 100
linguistic intelligence
personal intelligence
interpersonal
intrapersonal
logical-mathematical intelligence
kinesthetic intelligence
musical intelligence
spatial intelligence
In 1904, Alfred Binet and Théophile Simon were asked by the French Ministry of Education to create a practical and accurate method of assessing the children who could not profit from regular instruction. Binet choose to use a battery of tests, which made no pretense of measuring precisely any single faculty. Rather, it was aimed at evaluating the child general mental development with a heterogeneous group of tasks. Binet had noticed that children who had difficulties at school were very often late in other fields easily mastered by most pupils of the same age. It was their general development that was slow and Binet was more interested in the later than in any specific school subject.
The 30 tests on the 1905 scale ranged from very simple sensory tasks to complex verbal abstractions. The items were arranged by approximate level of difficulty instead of content. A rough standardization had been done with 50 normal children ranging from three to eleven years of age and several subnormal children as well. The notion of mental age originated directly from Binet's observation that, as they grow up, children can learn increasingly difficult concepts and ideas and do increasingly difficult things. This allowed Binet and Simon to order their tests according to the age level at which they were typically passed.
Wilhelm Stern quickly decided to express mental development as a ratio computed from the mental age (obtained from Binet's tests) that he then divided by the chronological age of the child. He obtained a number he called the MQ for Mental Quotient. Lewis Terman suggested multiplying the Mental Quotient by 100 to remove fractions and he created the Intelligence Quotient or IQ, which has survived to our days:
Psychologist Howard Gardner
Gardner, H. (1983). Frames of Mind: The theory of multiple intelligences. New York: Basic Books. Basic Books Paperback, Tenth Anniversary Edition with new introduction, New York: Basic Books, 1993.
dr hab. inż. Joanna Józefowska, prof. PP

14. General intelligence Charles Spearman G-factor
Intelligence is not a collection of various aptitudes but the integration of various aptitudes into a coherent whole.
Humans are smarter than computers because they can switch from Chess to Painting and see the connections between those fields, something that computers are completely unable to do.
Intelligence is at least as much into the links between our various aptitudes that into the various aptitudes themselves and it is a serious mistake to reduce intelligence to the aptitudes that support it.
Charles Spearman was puzzled by his discovery: "mental abilities of nearly all kinds are positively linked in the sense that if we are good at one thing, we are also likely to be good at others." If a person has a good vocabulary, there is a better than even chance that she has a good memory and that arithmetic is not a problem. Similarly, if a person is good at arithmetic, she probably has a better than average vocabulary or memory. These associations are not always true, but they are true on average and it is said that all our abilities are intercorrelated.
Spearman proposed the simplest possible explanation for this universal fact. Intelligence would consist of two kinds of factors: a single general factor, the G factor (or G) that would explain all the observed correlations, and numerous specific factors, s1, s2, ... that would account for the differences in test scores. Suddenly, it became clear why the IQ was such a good measure of mental growth and of general mental functioning: the IQ was an average where the specific factors would cancel each other out and let the G factor stand out! Spearman was very curious, would some specific tests be more "G loaded" than others and if yes, what would those tasks be? Only insight into the nature of such tasks would let him know whether he had found something trivial or worthwhile.
To understand better what Spearman did and what the G factor is, the following analogy with school may be useful. We are all familiar with schools /universities and with the concept of the bright student vs. the "not so bright" student. The GPA (Grade Point Average or overall mean) is calculated everywhere in the world to evaluate students on a single dimension: scholastic success. Schools and universities can be harsh or lenient but school subjects are always intercorrelated and a general factor of school success - the GPA - does exist. This general factor is correlated with absolutely all school subjects. Spearman, once he discovered the G factor, was like a school principal curious about which subjects would be the best summary of his entire curriculum (or that would be most correlated with the GPA). The only difference being that Spearman's "curriculum" was the entire range of human skills and abilities and that his "GPA" was the IQ.
To understand the nature of G was not easy, the tests or items with a high G loading (highly correlated with the IQ) were not similar at all at first glance, and neither were those with a low G loading (low correlation with the IQ). "All sorts of vehicles could carry G." The great discovery of Spearman, which has been refined ever since, was that the tests or items were loaded on G in direct proportion to the level of mental complexity involved to solve them ! The best measures of G are still those where one must compare and choose, analyze and synthesize, induce and deduce purposefully or discover structures and infer properly. In other words, judgment and invention (see previous definition of intelligence) are the two exact synonyms of the G factor.
Let's summarize: Binet invented the notion of mental age to describe the global level of children's cognitive development. This approach led to the notion of IQ, a kind of super-average of all our mental aptitudes. Spearman extracted the quintessence of the IQ, the G factor, by identifying the items loading it the most. The G factor turned out to be more, qualitatively, than the mere sum of the different elements involved in the original IQ tests: G was the ability to manage complexity, the essence of intelligence itself.
 Idiot-savants may have very high aptitudes in one field but have very low IQs, this is why they are considered "idiots".
dr hab. inż. Joanna Józefowska, prof. PP

15. General intelligence memory creativity imagination common sense
Intelligence and
memory
creativity
imagination
common sense
intuition
emotions
morality
dr hab. inż. Joanna Józefowska, prof. PP

16. Famous meeting
"Within ten years a digital computer will be the world's chess champion," Allen Newell said in 1957, "unless the rules bar it from competition."
The Dartmouth Seminar 1956
Dartmouth College: John McCarthy
Marvin Minsky
Claude Shannon
Nathaniel Rochester
Princeton: Trenchard More
IBM: Arthur Samuel
MIT: Ray Solomonoff
Oliver Selfridge
Carnegie Tech: Allen Newell
Herbert Simon
An eclectic array of academic and corporate scientists viewed the demonstration of the Logic Theorist at what became the Dartmouth Summer Research Project on Artificial Intelligence. The attendance list read like a present-day Who's Who in the field: John McCarthy, creator of the popular AI programming language LISP and director of Stanford University's Artificial Intelligence Laboratory; Marvin Minsky, leading AI researcher and Donner Professor of Science at M.I.T.; Claude Shannon, Nobel Prize-winning pioneer of information and AI theory, who was with Bell Laboratories.     By the end of the two-month conference, artificial intelligence had found its niche. Thinking machines and automata were looked upon as antiquated technologies. Researchers' expectations were grandiose, their predictions fantastic. "Within ten years a digital computer will be the world's chess champion," Allen Newell said in 1957, "unless the rules bar it from competition."
dr hab. inż. Joanna Józefowska, prof. PP

17. Artificial intelligence
Thinking humanly
Thinking rationally
Acting humanly
Acting rationally
Source: Russel S.J., Norvig P., Artificial intelligence - a modern approach, Prentice Hall 1995.
dr hab. inż. Joanna Józefowska, prof. PP

18. The Turing test ?
dr hab. inż. Joanna Józefowska, prof. PP

19. Criticism of the Turing test
The Test provides a guarantee not of intelligence but of culturally-oriented human intelligence (see French, Robert M.: Subcognition and the Limits of the Turing Test).
The test is limited to solving symbolic tasks, it is not possible to verify perception or manual abilities, although they reflect human intelligence.
But even if the second candidate was a real person from a different culture, she might be considered a 'machine' (i.e., not intelligent) because of certain questions she wouldn't be able to answer, or would answer in an unexpected way. For example, asking about the side of the road you drive on would be answered in different ways - that's a simple example, but one most people will understand. There are more subtle differences people might not be aware of, maybe a minor detail in everyday life you take for granted to be one way, while it is different in a different part of the world. Thus, the Turing Test shares its fate with early IQ tests the US Army used, and that immigrants usually failed because of their little knowledge of American culture.
dr hab. inż. Joanna Józefowska, prof. PP

20. Defense of the Turing test
The only standard allowing to discover intelligence without defining its „true” nature.
It ignores the problem of internal computer inference mechanism and its consciousness.
The natural advantages of „living” object is reduced by the interface.
dr hab. inż. Joanna Józefowska, prof. PP

21. Mental faculties flavour, test, intuition
dr hab. inż. Joanna Józefowska, prof. PP

22. Application domains of artificial intelligence
Natural language processing
Image recognition
Automated reasoning
Games
Expert systems
Automatic learning
Action planning and robotics
dr hab. inż. Joanna Józefowska, prof. PP

23. Cognitivism or connectionism? Weak or strong artificial intelligence?
dr hab. inż. Joanna Józefowska, prof. PP

24. Conectionist model Cognitivist model How does a human brain work?
How do humans solve problems?
Big number of identical simple units
Distributed and parallel processing
Failure resistance
Symbolic knowledge representation
Inference mechanism
Complexity of learning
Complexity of search
dr hab. inż. Joanna Józefowska, prof. PP

25. Physical symbol system hypothesis 1976 - Allen Newell, Herbert A. Simon*)
A physical symbol system consists of a set of entities, called symbols, which are physical patterns that can occur as components of another type of entity called an expression (or symbol structure).
Hypothesis:
A physical symbol system has the necessary and sufficient means for general intelligent action.
*)Carnegie Tech - now Carnegie Mellon University
dr hab. inż. Joanna Józefowska, prof. PP

26. Knowledge representation
The process and the result of formalization of knowledge in such a way, that it can be used automatically for problem solving.
Search
General technique for problem solving consisting in systematic exploration of all consecutive and alternative steps in the problem solving process.
dr hab. inż. Joanna Józefowska, prof. PP

27. Automated reasoning Alfred N. Whitehead 1861-1947 Bernard Russel
dr hab. inż. Joanna Józefowska, prof. PP

28. 1931 –incompleteness theorem
Let NT be a set of axioms of the number theory. If theory T(NT) is sound then it is incomplete*).
*) Proof in:
[1] Smullyan R., What is the name of this book? – The riddle of Draculla and other logical puzzles. Prentice-Hall, 1978.
[2] Mendelson E., Introduction to mathematical logic, Wyd. 4, Chapman and Hall, 1997.
Kurt Gödel
dr hab. inż. Joanna Józefowska, prof. PP

29. Logic Theorist - 1956 Allen Newell Herbert Simon
Herbert Simon
Newell and Simon opracowali własny język oparty na reprezentacji listowej: IPL. Nie mieli kompilatora i translację na kod maszynowy przeprowadzali „ręcznie”. Aby uniknąć błędów pracowali niezależnie i równolegle, a następnie czytali sobie nawzajem ciągi zer i jedynek i zapisywali każdą instrukcję dopiero po jej „uzgodnieniu”, gdy żaden z nich nie miał wątpliwości co do jej poprawności. (źródło: Russel i Norvig s. 17 footnote).
dr hab. inż. Joanna Józefowska, prof. PP

30. First Order Logic (FOL)
A = Z  F  P  S  {(, ), ,}
A – set of symbols
Z - variables x1, x2, ...
F – function symbols: F1n, F2n, ...
P – predicate symbols: P1n, P2n, ...
S – logical symbols {, , , , , , }
dr hab. inż. Joanna Józefowska, prof. PP

31. Deduction Inference method based on modus ponens
dr hab. inż. Joanna Józefowska, prof. PP

32. Theory (A1) man(X)  die(X) (A2) man(sokrates) Theorem: die(sokrates)
Dowód:
1. X/sokrates in A1
(A1’) man(sokrates)  die(sokrates)
2. modus ponens
man(sokrates)
, man(sokrates)  die(sokrates)
die(sokrates)
dr hab. inż. Joanna Józefowska, prof. PP

33. Operations in Logic Theorist
Substitution: any variable may be substituted by an expresion.
e.g. in (ØAÚ B) Û (A Þ B) we substitute ØA for B (ØA Ú ØA) Û (AÞØA) (*)
Replacement: an operator can be replaced by a definition.
np. w wyrażeniu (ØA Ú ØA) ÞØA zastępujemy operator Ú jego definicją (*)
(A Þ ØA) Þ ØA
Modus ponens (reguła odrywania):
[(AÞ B)Ù A] B
dr hab. inż. Joanna Józefowska, prof. PP

34. Logic Theorist - summary
Newella, Simona and Shawa 1956
Proved theorems from the first chapter of Principia Mathematica
Knowledge representation: FOL
Inference: deduction
Comparison of expressions: unification
Problems: complexity
For one of the equations, Theorem 2.85, the Logic Theorist surpassed its inventors' expectations by finding a new and better proof.
dr hab. inż. Joanna Józefowska, prof. PP