Computability - PowerPoint PPT Presentation

About This Presentation
Title:

Computability

Description:

Computability Kolmogorov-Chaitin-Solomonoff. Other topics. Homework: Prepare presentations. Information Shannon definition: series of (binary) choices. – PowerPoint PPT presentation

Number of Views:33
Avg rating:3.0/5.0
Slides: 14
Provided by: Jeani200
Category:

less

Transcript and Presenter's Notes

Title: Computability


1
Computability
  • Kolmogorov-Chaitin-Solomonoff. Other topics.
  • Homework Prepare presentations.

2
Information
  • Shannon definition series of (binary) choices.
    Information in measured in bits.
  • Computability definition Let x be a binary
    string. A minimal description of x, d(x) is the
    shortest string,ltM,wgt, where Turing Machine M on
    input w halts with x on the tape. The descriptive
    complexity (aka Kolmogorov or Kolmogorov-Chaitin
    complexity) is K(x) d(x)
  • the length of this shortest string.
  • Note there may be more than one.

3
Informally
  • Suppose we have a string consisting of 100 groups
    of '0110'. This is a description and it seems
    like it would be shorter than writing out the
    whole string.
  • The formal description includes the TM (or
    program) that knows what to do with 100 groups of
    something. The full minimal description
    consists of this TM plus an encoding of 100
    groups, 0110.
  • Program that takes description and produces
    string.
  • This definition requires a TM for even the
    definition that is the whole string.

4
Common example
  • Mandelbrot fractals are very intricate patterns
    and yet can be produced (re-produced?) by simple
    computer programs.

5
Claim
  • There exists a constant c, such that for all x,
    K(x)lt x c.
  • intuitively take a fixed TM that halts
    immediately. Then it halts with the string x on
    it.
  • There exists a constant c, such that for all x
    and y, K(xy)lt2K(x) K(y)c

6
Incompressible strings
  • Definitions a string x is c-compressible if
    K(x)ltx-c.
  • If x is not c-compressible, x is incompressible
    by c.
  • If x is incompressible by 1, x is incompressible.

7
Incompressible strings exist!
  • The number of strings of length n is greater than
    the number of descriptions of length less than n.
    So, some string of length n is not described by
    any description of length less than n.

8
K(x) is not computable!
  • and not because the definition is tied to any
    one of several models of computing
  • See http//en.wikipedia.org/wiki/Kolmogorov_comple
    xity It is a Halting Problem type of proof.

9
Berry Paradox
  • Let n be the smallest positive integer that
    cannot be defined in fewer than twenty English
    words.
  • oops!
  • Relates also to Godel incompleteness results.

10
Recursion Theorem
  • also called fixed point theorem.
  • There exists a program that prints itself.
  • General strategy create constant that represents
    working part of program and write program that
    prints out constant.
  • See Logo example in Shai lectures.
  • c example next slidehttp//igoro.com/archive/ho
    w-to-write-a-self-printing-program/

11
c example
  • May need to force to be in two lines
  • class Pstatic void Main()var S"class Pstatic
    void Main()var S101System.Console
  • .Write(S,S,'1')"System.Console.Write(S,S,'
    "')

12
finish Breaking the code
  • Discussion

13
Homework
  • Prepare presentations!
Write a Comment
User Comments (0)
About PowerShow.com