Deductive and inductive methods for program synthesis - PowerPoint PPT Presentation

About This Presentation
Title:

Deductive and inductive methods for program synthesis

Description:

Deductive and inductive methods for program synthesis. Jelena Sanko, Jaan Penjam ... pi= R1; R2; R3; R4; R5; ... ;R6. Square root fitness measure of the program ... – PowerPoint PPT presentation

Number of Views:188
Avg rating:3.0/5.0
Slides: 16
Provided by: jele7
Category:

less

Transcript and Presenter's Notes

Title: Deductive and inductive methods for program synthesis


1
Deductive and inductive methods for program
synthesis
  • Jelena Sanko, Jaan Penjam
  • Institute of Cybernetics

October 29, 2005
2
Agenda
  • Program construction
  • Inductive program construction
  • Experiments
  • Conclusions

3
Program construction
4
Problem Statement
x
5
Scoring of programs
s - measure of fitness
pi?R1 R2 R3 R4 R5 ... R6?
Square root fitness measure of the program p
6
Program synthesis as optimization
s
z0
z
1
h
P?R2 R3 R5 R6 R2 R6?
7
Coding of programs
z 0,235627
h
p?R2 R3 R5 R6 R2 R6?
8
PR6
Coding of programs (2)
M(P) - State Transition Machine
computational model
9
Coding of programs (3)
2
x
,
3
1
xy
xz
2
4
,
1
3
xyz
vxz
5
45
2
,
1
4
uvxz
vxyz
2
,
1
5
0.33
0.66
uvxyz
45
6
tuvxyz
10
Function to be optimized (f(z))
s
1
z
11
Optimization technique
-Differential Evolution,
Rainer Storn and Kenneth Price
DE is a method for finding extreme points of
real-valued multi-modal functions
DE is a heuristic method that can be used for
optimization of non-differentiable functions in
continuous spaces
The convergence rate of the floating-point
encoded DE is more than 10 times higher than the
convergence rate of the traditional binary
encoded GA
12
Experimental Data
I(R1)yln(x) I(R2)yexp(-x) I(R3)zx3
6x I(R4)v2(x2z) I(R5)(u,z)(v6,v/2) I
(R6)v0.16yu/7.854
P(0.75,1.8710-07), (0.70, 1.0410-07) (0.65,5
.4410-08), (0.60,2.6110-08) (0.55,1.1710-08)
,(0.50, 3.4010-09)
P.M. Murphy and D.W. Aha, Uci repository of
machine learning databases
13
Experimental Data (2)
I(R1)yln(x) I(R2)yexp(-x) I(R3)zx3
6x I(R4)v2(x2z) I(R5)(u,z)(v6,v/2) I
(R6)v0.16yu/7.854
The optimal solutions like 0.37362, 0.40419,
0.795112 correspond to the following sequence
of relations R2R3R4R5R4R5R4R5R6
The optimal solution of the problem obtained by
the DE method corresponds to the proposed by Knut
Angström functional relation
where
14
Conclusions
  • The inductive approach to program construction
    can promote control over construction process and
    take an appropriate decision when several
    solutions are available
  • IPS uses interpretation of relations and a
    fitting function
  • IPS considers all programs and selects the best
  • The research is in progress...
  • IPS strategy need to be checked in practice and
    extension to high order constraints is required

15
?
Write a Comment
User Comments (0)
About PowerShow.com