comparison toolboxes/FullBNT-1.0.7/bnt/CPDs/@boolean_CPD/boolean_CPD.m @ 0:e9a9cd732c1e tip

first hg version after svn
author wolffd
date Tue, 10 Feb 2015 15:05:51 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:e9a9cd732c1e
1 function CPD = boolean_CPD(bnet, self, ftype, fname, pfail)
2 % BOOLEAN_CPD Make a tabular CPD representing a (noisy) boolean function
3 %
4 % CPD = boolean_cpd(bnet, self, 'inline', f) uses the inline function f
5 % to specify the CPT.
6 % e.g., suppose X4 = X2 AND (NOT X3). Then we can write
7 % bnet.CPD{4} = boolean_CPD(bnet, 4, 'inline', inline('(x(1) & ~x(2)'));
8 % Note that x(1) refers pvals(1) = X2, and x(2) refers to pvals(2)=X3.
9 %
10 % CPD = boolean_cpd(bnet, self, 'named', f) assumes f is a function name.
11 % f can be built-in to matlab, or a file.
12 % e.g., If X4 = X2 AND X3, we can write
13 % bnet.CPD{4} = boolean_CPD(bnet, 4, 'named', 'and');
14 % e.g., If X4 = X2 OR X3, we can write
15 % bnet.CPD{4} = boolean_CPD(bnet, 4, 'named', 'any');
16 %
17 % CPD = boolean_cpd(bnet, self, 'rnd') makes a random non-redundant bool fn.
18 %
19 % CPD = boolean_CPD(bnet, self, 'inline'/'named', f, pfail)
20 % will put probability mass 1-pfail on f(parents), and put pfail on the other value.
21 % This is useful for simulating noisy boolean functions.
22 % If pfail is omitted, it is set to 0.
23 % (Note that adding noise to a random (non-redundant) boolean function just creates a different
24 % (potentially redundant) random boolean function.)
25 %
26 % Note: This cannot be used to simulate a noisy-OR gate.
27 % Example: suppose C has parents A and B, and the
28 % link of A->C fails with prob pA and the link B->C fails with pB.
29 % Then the noisy-OR gate defines the following distribution
30 %
31 % A B P(C=0)
32 % 0 0 1.0
33 % 1 0 pA
34 % 0 1 pB
35 % 1 1 pA * PB
36 %
37 % By contrast, boolean_CPD(bnet, C, 'any', p) would define
38 %
39 % A B P(C=0)
40 % 0 0 1-p
41 % 1 0 p
42 % 0 1 p
43 % 1 1 p
44
45
46 if nargin==0
47 % This occurs if we are trying to load an object from a file.
48 CPD = tabular_CPD(bnet, self);
49 return;
50 elseif isa(bnet, 'boolean_CPD')
51 % This might occur if we are copying an object.
52 CPD = bnet;
53 return;
54 end
55
56 if nargin < 5, pfail = 0; end
57
58 ps = parents(bnet.dag, self);
59 ns = bnet.node_sizes;
60 psizes = ns(ps);
61 self_size = ns(self);
62
63 psucc = 1-pfail;
64
65 k = length(ps);
66 switch ftype
67 case 'inline', f = eval_bool_fn(fname, k);
68 case 'named', f = eval_bool_fn(fname, k);
69 case 'rnd', f = mk_rnd_bool_fn(k);
70 otherwise, error(['unknown function type ' ftype]);
71 end
72
73 CPT = zeros(prod(psizes), self_size);
74 ndx = find(f==0);
75 CPT(ndx, 1) = psucc;
76 CPT(ndx, 2) = pfail;
77 ndx = find(f==1);
78 CPT(ndx, 2) = psucc;
79 CPT(ndx, 1) = pfail;
80 if k > 0
81 CPT = reshape(CPT, [psizes self_size]);
82 end
83
84 clamp = 1;
85 CPD = tabular_CPD(bnet, self, CPT, [], clamp);
86
87
88
89 %%%%%%%%%%%%
90
91 function f = eval_bool_fn(fname, n)
92 % EVAL_BOOL_FN Evaluate a boolean function on all bit vectors of length n
93 % f = eval_bool_fn(fname, n)
94 %
95 % e.g. f = eval_bool_fn(inline('x(1) & x(3)'), 3)
96 % returns 0 0 0 0 0 1 0 1
97
98 ns = 2*ones(1, n);
99 f = zeros(1, 2^n);
100 bits = ind2subv(ns, 1:2^n);
101 for i=1:2^n
102 f(i) = feval(fname, bits(i,:)-1);
103 end
104
105 %%%%%%%%%%%%%%%
106
107 function f = mk_rnd_bool_fn(n)
108 % MK_RND_BOOL_FN Make a random bit vector of length n that encodes a non-redundant boolean function
109 % f = mk_rnd_bool_fn(n)
110
111 red = 1;
112 while red
113 f = sample_discrete([0.5 0.5], 2^n, 1)-1;
114 red = redundant_bool_fn(f);
115 end
116
117 %%%%%%%%
118
119
120 function red = redundant_bool_fn(f)
121 % REDUNDANT_BOOL_FN Does a boolean function depend on all its input values?
122 % r = redundant_bool_fn(f)
123 %
124 % f is a vector of length 2^n, representing the output for each bit vector.
125 % An input is redundant if there is no assignment to the other bits
126 % which changes the output e.g., input 1 is redundant if u(2:n) s.t.,
127 % f([0 u(2:n)]) <> f([1 u(2:n)]).
128 % A function is redundant it it has any redundant inputs.
129
130 n = log2(length(f));
131 ns = 2*ones(1,n);
132 red = 0;
133 for i=1:n
134 ens = ns;
135 ens(i) = 1;
136 U = ind2subv(ens, 1:2^(n-1));
137 U(:,i) = 1;
138 f1 = f(subv2ind(ns, U));
139 U(:,i) = 2;
140 f2 = f(subv2ind(ns, U));
141 if isequal(f1, f2)
142 red = 1;
143 return;
144 end
145 end
146
147
148 %%%%%%%%%%
149
150 function [b, iter] = rnd_truth_table(N)
151 % RND_TRUTH_TABLE Construct the output of a random truth table s.t. each input is non-redundant
152 % b = rnd_truth_table(N)
153 %
154 % N is the number of inputs.
155 % b is a random bit string of length N, representing the output of the truth table.
156 % Non-redundant means that, for each input position k,
157 % there are at least two bit patterns, u and v, that differ only in the k'th position,
158 % s.t., f(u) ~= f(v), where f is the function represented by b.
159 % We use rejection sampling to ensure non-redundancy.
160 %
161 % Example: b = [0 0 0 1 0 0 0 1] is indep of 3rd input (AND of inputs 1 and 2)
162
163 bits = ind2subv(2*ones(1,N), 1:2^N)-1;
164 redundant = 1;
165 iter = 0;
166 while redundant & (iter < 4)
167 iter = iter + 1;
168 b = sample_discrete([0.5 0.5], 1, 2^N)-1;
169 redundant = 0;
170 for i=1:N
171 on = find(bits(:,i)==1);
172 off = find(bits(:,i)==0);
173 if isequal(b(on), b(off))
174 redundant = 1;
175 break;
176 end
177 end
178 end
179