1 / 109

LISP (Lecture Note #7)

LISP (Lecture Note #7). 인공지능 이복주 단국대학교 컴퓨터공학과. Outline. What is LISP Variables and Functions Data Types SETF Math Operations S-expressions CONS Cell …. What is LISP?. A LISt Processing language The basic data structure is linked list A functional programming language

emmett
Télécharger la présentation

LISP (Lecture Note #7)

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LISP(Lecture Note #7) 인공지능 이복주 단국대학교 컴퓨터공학과

  2. Outline • What is LISP • Variables and Functions • Data Types • SETF • Math Operations • S-expressions • CONS Cell • … Slide made by Bogju Lee

  3. What is LISP? • A LISt Processing language • The basic data structure is linked list • A functional programming language • Each expression in LISP is a function that returns a value • An interpretive language • Prompt at which you type commands. • Type-eval-print loop • Two things to type: variables or functions • Running LISP programs involves interacting with the LISP interpreter • LISP programs can also be compiled

  4. Why LISP? Why LISP for AI programming? • LISP supports symbol manipulation better • Symbols are the basic entity of the language • The interpretive nature makes it easy to “try a new idea’’ and prototype • Functions are in the form of linked lists • Easier to write learning programs • Historical Reason • Most AI programs in the U.S. have been developed in LISP • However, most AI programs in Europe have been developed in PROLOG

  5. Variables and Functions • Variables • Non-case sensitive • No length limit • When typed at prompt, print the value “bounded”. • Functions • Form of a list of things enclosed in parentheses. • (Name_of_the_function argument1 argument2 …) • Evaluation order: argument1, argument2, .. and then call the function.

  6. Data Types • Integers, floating-point numbers, strings, and characters • Symbols • Kind of strings, but not surrounded by double-quote. • Stored in efficient internal data structures (hash tables) so • LISP can quickly determine when two symbols are identical. • Variable vs. Symbol • To differentiate, use a single-quote for symbol. • > A • 23 • > ‘A • A

  7. Special Symbols • NIL represents an empty list. • NIL is a terminator of a list. • A list is usually built by inserting its elements into NIL in the reverse order . • NIL can also represent “false”. • The special symbol representing “true” is T.

  8. SETF • (SETF <variable> <expression> • Evaluate the expression and set the value to the variable • Example • > (setf a 3) ; comment start with ‘;’ • 3 • > (setf b “foo”) ; a string • “foo” • > (setf c ‘bar) ; a symbol • BAR ; converted to upper case • > (setf d b) • “foo”

  9. Math Operations • +, -, *, / • FLOAT, ROUND, SQRT, EXPT, LOG, EXP

  10. S-expressions • A list is an S-expression. • Nested list is also an S-expression. • (1 2 (3 4) 5))

  11. CONS Cell • CONS cell: dynamically-allocated piece of memory in the LISP image • Automatic garbage collection • A CONS cell contains two pointer: CAR and CDR. • The CAR pointer usually points to a value (e.g., a symbol, or a number). • The CDR pointer usually points to the next CONS cell in a linked list. CDR CDR CAR CAR

  12. List • A linked list always terminates by having the CDR of the last CONS cell pointing to a special symbol: NIL. • Example: (John loves Mary) NIL John loves Mary

  13. NIL Mary CONS • The function CONS returns a newly created CONS cell. • The function takes two arguments. • The CAR part of the new cell points to the first argument. • The CDR part of the new cell points to the second argument. • Example: (CONS ‘Mary NIL) returns • (MARY)

  14. CONS (2) • Using CONS to insert an item to the front of a list • (CONS <item to be inserted> <a list>) • Example: (CONS 'John (CONS 'loves (CONS 'Mary NIL))) • returns • First CONS call returns (MARY) • Second CONS call returns (LOVES MARY) • Third CONS call returns (JOHN LOVES MARY)

  15. NIL John loves NIL play tennis Nested Sublists • A sublist is a list pointed by a CAR pointer of a CONS cell • Example: ( John loves ( play tennis ) )

  16. CONS (3) • Using CONS to insert sublists • When CONS insert a list into another list. The former becomes a sublist of the latter. • Example: • > (CONS '(Amazon BN) '(sells books on Internet)) • ((AMAZON BN) SELLS BOOKS ON INTERNET)

  17. CAR, CDR • List access • (CAR <a list>) returns the first element of the list. • (CDR <a list>) returns the remaining list (i.e., everything except the first element). • Example: (CAR '(John loves Michelle)) returns • JOHN • Example: (CDR '(John loves Michelle) ) returns • (LOVES MICHELLE)

  18. LIST • Other list creating functions • (LIST <item1> <item2> ... <itemN>) returns a list • whose first element is <item 1>, second element is • <item 2>, ...., Nth element is <itemN>. • Example: • > (LIST 'John 'loves ’Michelle) • (JOHN LOVES MICHELLE) • > (LIST 'John 'loves (LIST `AI `research)) • (JOHN LOVES (AI RESEARCH)) • > (SETF A 3) • > (LIST A (LIST 1 2) ‘foo) • (3 (1 2) FOO)

  19. APPEND • Concatenating lists • (APPEND <list 1> <list 2>) returns a list that is the result of concatenating <list 1> and <list 2>. • > (APPEND '(Orange Apple) '(Grape)) • (ORANGE APPLE GRAPE) • APPEND can also be used to concatenate more than two lists. • > (APPEND '(play) '(tennis) '(football baseball) ) ) ) • (PLAY TENNIS FOOTBALL BASEBALL)

  20. Other List Functions • LENGTH • > (length ‘(1 2 (3 4) 5)) • 4 • FIRST • > (first ‘(a b c)) • A • SECOND • > (second ‘(a b c)) • B • REST • > (rest ‘(a b c)) • ‘(B C) • NTH • > (nth 2 ‘(a b c)) • C • REVERSE • > (reverse ‘(a b c)) • (C B A)

  21. Evaluation • LISP executes/interprets an expression through an evaluation procedure. • Example: • (+ 3 5) evaluates to ==> 8 • (CONS ‘A NIL) evaluates to ==> (A) • 3 evaluates to ==> 3 • ‘A evaluates to ==> A • (> 5 3) evaluates to ==> T

  22. Evaluation Rules • A number evaluates to itself • A symbol evaluates to its value. • SETQ assigns a value to a symbol • A list is evaluated by • treating the first element as a function • evaluating each arguments of the function in a left-to-right order • An expression preceeded by a quote symbol ‘ evaluates to the expression itself.

  23. Quote • Quote symbol ‘ is a short hand for a function QUOTE • (QUOTE <arg>) • QUOTE is a special function that prevents LISP from evaluating its argument. • QUOTE returns the argument literately. • Example: (quote (dummy-fn 2)) • ==> (DUMMY-FN 2)

  24. SETQ • Assignment and binding • A symbol (variable) can be assigned a value (called its binding) using SETQ. • (SETQ <symbol-name> <anything>) • Example: (SETQ A ‘(A B C)) ==> (A B C) A evaluates to ==> (A B C) • Evaluating a symbol that does not have a value assigned (i.e., no binding) causes error • Ex: B evaluates to ==> Error: Unbound Variable

  25. Change Binding • All other functions do NOT change the bindings • In particular, CAR and CDR are non-destructive. • > (setq my-friends ‘(Superman Batman Robin) ) • (Superman Batman Robin) • > (car (cdr my-friends)) • Batman • > my-friends • (Superman Batman Robin)

  26. Change Binding (2) • CONS does not change bindings either. • > (CONS ‘J-Bond my-friends) • (J-Bond Superman Batman Robin) • > my-friends • (Superman Batman Robin) • > (setq my-friends (CONS ‘J-Bond my-friends) ) • (J-Bond Superman Batman Robin) • > my-friends • (J-Bond Superman Batman Robin)

  27. DEFUN • A function is defined using DEFUN • (DEFUN <fn-name> (<arg1> ...<argK>) <exp1> ... <expN> ) • All arguments are passed by value. • The body of the function may contain any number of expressions (i.e., function calls) • The function returns the value returned by the last expression.

  28. DEFUN (2) • >(defun range (n) • (cond ((not (numberp n)) 'not-number) • ((< n 0) 'negative) • ((> n 0) 'positive) • (T 'zero))) • > (range 52.03) • POSITIVE

  29. DEFUN (3) • (defun square (x) • (times x x)) • (defun add-friend (new-friend friends) • (cons new-friend friends) ) • > (square 5) • 25 • > (add-friend ‘B-Gates my-friends) • (B-Gates J-Bond Superman Batman Robin)

  30. Symbol and Function with the Same Name • Changing a symbol’s binding does not change its function definition • >(setq cons ‘list) • LIST • >(setq b (cons cons nil)) • (LIST) • > cons • LIST • > (cons ‘cons b) • (CONS LIST)

  31. Predicates • Functions that return ``true’’ (i.e., T) or ``false’’ (i.e., NIL). • type-testing predicates • (NULL <item>) returns T if <item> is NIL (empty list), otherwise NIL. • (LISTP <item>) returns T if <item> is a list, otherwise NIL. • (ATOM <item>) returns T if <item> is an atom (i.e., a symbol, a number, or NIL). • (NUMBERP <item>) returns T if <item> is a number

  32. EQ, EQUAL • Equality-testing Predicates: EQ, =, STRING=, EQUAL • (EQ <item1> <item2>) returns T if two arguments are two identical atoms (symbols and characters). • > (EQ ‘(Banana Apple) ‘(Banana Apple)) • NIL • (= <number1> <number2>) • > (= 5 (/ 10 2)) • T • (STRING= <string1> <string2>) • (EQUAL <item1> <item2>) returns T if two arguments are identical atoms or identical lists (S-expressions). • > (EQUAL ‘(Banana Apple) ‘(Banana Apple)) • T • > (EQUAL (LIST ‘a ‘b) (list ‘b ‘a)) • NIL

  33. MEMBER • (MEMBER <item> <list>) returns T if <item> is an atom that is a top-level element of <list>. • > (setq my-friends ‘(MacGyver (Batman Robin)) ) • > (member ‘MacGyver my-friends) • T • > (member ‘Batman my-friends) • NIL • > (member ‘(Batman Robin) my-friends) • NIL

  34. Summary • What is LISP • Variables and Functions • Data Types • SETF • Math Operations • S-expressions • CONS Cell • … Slide made by Bogju Lee

  35. Outline • IF, PROGN • COND • AND, OR, NOT • Recursive Functions • LET • FORMAT • EVAL • APPLY • FUNCALL • MAPCAR • REMOVE Slide made by Bogju Lee

  36. IF, PROGN • Conditional Expression: IF, PROGN • (IF <condition> <expression1> [<expression2>]) • Takes the first expression if the condition is non-nil (true) • Optional second expression is taken if the condition is nil (false). • > (if (< 2 3) ‘A (+ 2 3)) • A • (PROGN <expression1> <expression2> ..) • Combine multiple forms to evaluate as a single consequence • > (if (= a b) ; suppose A=5 and B=5 • (progn • (format t "A equals B.~%") ; ~% means new line. • a) ; trivial case • (/ (+ a b) 2)) ; print the average • A equals B. • 5

  37. COND • COND is an N-branch conditional expression • (COND ( <test1> <exp11> ... <exp1L> ) • ( <test2> <exp21> ... <exp2M> ) • ... • ( <testK> <expK1> ... <expKN> ) ) • Each test is evaluated sequentially until a test returns true. • Expressions following that test will be executed. • COND returns the value returned by the last expression associated with the test

  38. COND - Example • (defun select-character (enemy) • (cond ( (eq enemy ‘Penguin) • ‘Batman) • ( (eq enemy ‘Catwoman) • ‘J-Bond ) • ( (eq enemy ‘Black-Knight) • ‘(White-Knight King-Arthur) ) • ) • )

  39. COND with T Condition • > (setf a –3) • > (cond ((< a 0) ‘negative) • ((> a 0) ‘positive) • (T ‘zero)) • NEGATIVE • (defun select-character (enemy) • (cond ( (eq enemy ‘Penguin) • ‘Batman) • ( (eq enemy ‘Catwoman) • ‘J-Bond ) • ( (eq enemy ‘Black-Knight) • ‘(White-Knight King-Arthur ) ) • ( T ; for all other enemies • ‘SuperMan) ; ask Superman for help • ) • )

  40. AND, OR, NOT • Binding more complex conditions • Simple tests can be combined into a complex one using AND, OR, NOT. • (AND <test1> <test2> ... <testn>) evaluates tests in a left-to-right order. • It returns NIL when it encounters the first test that evaluates to NIL (remaining tests are not evaluated). • If all the tests return non-NIL, AND returns the value of the last test (i.e., testn). • (Defun safe-caar ( L ) • (and (listp L) • (listp (car L)) • (car (car L)) • ) )

  41. AND • > (safe-caar ‘A) • NIL • > (safe-caar ‘( ( A ) ) ) • A • > (caar ‘A) • Error!

  42. OR and NOT • (OR <test1> <test2> ... <testn>) evaluates tests in a left-to-right order. • It returns T when it encounters the first test that evaluates to T (remaining tests are not evaluated). If all the tests return NIL, OR returns NIL. • (NOT <exp>) returns T if <exp> is NIL, returns NIL if <exp> is non-NIL. • > (NOT (NULL ‘(A B C) ) ) • T

  43. Using AND, OR in COND • (defun Evaluate-feeling ( sentence ) • (cond ( (OR (member ‘hate sentence) • (member ‘dislike sentence)) • ‘hatred) • ( (AND (member ‘I sentence) • (member ‘feel sentence) ) • ‘self-centered ) • ( T • ‘happy) • ) ; end of cond • ) ; end of defun

  44. Martin and the Dragon • Martin: Are there enemies in the list of attendees of King Arthur’s party tonight? • (Micky SnowWhite Bishop BlackKnight WhiteKnight RoboCop SirLancelot Modred Magician) • Dragon: I will only tell you whether the first one in the list is an enemy or not. • What should Martin do?

  45. Recursive Functions • Functions that call themselves • A LISP version of Martin’s solution • (defun find-enemy ( guest-list ) • (cond ( (member (car guest-list) enemies) • T ) ; Yes, there is an enemy • ( T • (find-enemy (cdr guest-list) ) ) • ) ; end of cond • ) ; end of defun

  46. Factorial Example • N! = N*(N-1)! • >(defun fact (n) (* n (fact (- n 1)))) • FACT • This is actually incomplete! • Consider the termination condition. • >(defun fact (n) • (if (= n 1) 1 (* n (fact (- n 1))))) • FACT • >(fact 5) • 120

  47. Member Example • Recursion is especially useful for list processing. • Many problems can be thought of recursively as applying to either the first element of a list or the rest. • Determine whether some symbol occurs in a list. • >(defun our-member (sym lst) • (cond ((null lst) nil) ; return false if empty list • ((eq sym (first lst)) T) ; we can stop here • (T (our-member sym (rest lst))))) ; ow, recursion • OUR-MEMBER • >(our-member 'b '(a b c)) • T

  48. Fixing the Infinite Recursion • (defun find-enemy ( guest-list ) • (cond ( (null guest-list) • NIL ) ; No, there are no enemies. • ( (member (car guest-list) enemies) • T ) ; Yes, there is an enemy • ( T • (find-enemy (cdr guest-list) ) ) • ) ; end of cond • ) ; end of defun

  49. Martin and the Dragon • Martin: Can you tell me the names of those enemies that are in the attendee list? • Dragon: (yawn) I will only tell you whether the first attendee in the list is an enemy or not.

  50. Building Lists Using Recursion • (defun identify-enemy ( guest-list ) • (cond ( (null guest-list) • NIL ) ; Reached the end of the guest list. • ( (member (car guest-list) enemies) • (car guest-list)) ; The first guest is an enemy. • ( T • (identify-enemy (cdr guest-list) ) ) • ) ; end of cond • ) ; end of defun

More Related