diogo/yap-6.3
Archived
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

improve learning for CLP(BN).

Browse Source
This commit is contained in:
Vítor Santos Costa 2011-11-30 13:04:13 +00:00
parent 0b81d99803
commit 2f6ec5ff59
6 changed files with 156 additions and 78 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
packages/CLPBN/clpbn.yap
View File

@ -11,6 +11,8 @@
clpbn_init_solver/5,
clpbn_run_solver/4,
clpbn_init_graph/1,
probability/2,
conditional_probability/3,
op( 500, xfy, with)]).
:- use_module(library(atts)).
@ -113,6 +115,8 @@
solver(ve).
em_solver(ve).
:- meta_predicate probability(:,-), conditional_probability(:,:,-).
%output(xbif(user_error)).
%output(gviz(user_error)).
output(no).
@ -468,3 +472,49 @@ clpbn_finalize_solver(State) :-
arg(Last, State, Info),
finalize_bp_solver(Info).
clpbn_finalize_solver(_State).
probability(Goal, Prob) :-
findall(Prob, do_probability(Goal, [], Prob), [Prob]).
conditional_probability(Goal, ListOfGoals, Prob) :-
\+ ground(ListOfGoals), !,
throw(error(ground(ListOfGoals),conditional_probability(Goal, ListOfGoals, Prob))).
conditional_probability(Goal, ListOfGoals, Prob) :-
findall(Prob, do_probability(Goal, ListOfGoals, Prob), [Prob]).
do_probability(Goal, ListOfGoals, Prob) :-
evidence_to_var(Goal, C, NGoal, V),
call_residue(run( ListOfGoals, NGoal), Vars), !,
match_probability(Vars, C, V, Prob).
run(ListOfGoals,Goal) :-
do(ListOfGoals),
call(Goal).
do(M:ListOfGoals) :-
do(ListOfGoals, M).
do([]).
do([], _M).
do(G.ListOfGoals, M) :-
M:G,
do(ListOfGoals, M).
evidence_to_var(M:Goal, C, M:VItem, V) :- !,
evidence_to_var(Goal, C, VItem, V).
evidence_to_var(Goal, C, VItem, V) :-
Goal =.. [L|Args],
variabilise_last(Args, C, NArgs, V),
VItem =.. [L|NArgs].
variabilise_last([Arg], Arg, [V], V).
variabilise_last([Arg1,Arg2|Args], Arg, Arg1.NArgs, V) :-
variabilise_last(Arg2.Args, Arg, NArgs, V).
match_probability([p(V0=C)=Prob|_], C, V, Prob) :-
V0 == V,
!.
match_probability([_|Probs], C, V, Prob) :-
match_probability(Probs, C, V, Prob).

10
packages/CLPBN/clpbn/bp.yap
View File

@ -58,8 +58,7 @@ init_bp_solver(_, AllVars, _, (BayesNet, DistIds)) :-
sort(DistIds0, DistIds),
%(num_bayes_nets(0) -> writeln(vars:VarsInfo) ; true),
%(num_bayes_nets(0) -> writeln(dists:DistsInfo) ; true),
create_network(VarsInfo, BayesNet),
get_extra_vars_info(AllVars, ExtraVarsInfo).
create_network(VarsInfo, BayesNet).
%set_extra_vars_info(BayesNet, ExtraVarsInfo).
@ -99,17 +98,10 @@ get_evidence(_V, -1). % no evidence !!!
get_extra_vars_info([], []).
get_extra_vars_info([V|Vs], [v(VarId, Label, Domain)|VarsInfo]) :-
get_atts(V, [id(VarId)]), !,
writeln(k:V),
clpbn:get_atts(V, [key(Key),dist(DistId, _)]),
writeln(j:Key),
term_to_atom(Key, Label),
writeln(i:Label),
get_dist_domain(DistId, Domain0),
writeln(a:Domain0),
numbers2atoms(Domain0, Domain),
writeln(b:Domain),
get_extra_vars_info(Vs, VarsInfo).
get_extra_vars_info([_|Vs], VarsInfo) :-
get_extra_vars_info(Vs, VarsInfo).

34
packages/CLPBN/clpbn/table.yap
View File

@ -12,6 +12,8 @@
clpbn_tableallargs/1,
clpbn_table_nondet/1,
clpbn_tabled_clause/2,
clpbn_tabled_clause_ref/3,
clpbn_tabled_retract/2,
clpbn_tabled_abolish/1,
clpbn_tabled_asserta/1,
clpbn_tabled_assertz/1,
@ -31,6 +33,8 @@
:- meta_predicate clpbn_table(:),
clpbn_tabled_clause(:.?),
clpbn_tabled_clause_ref(:.?,?),
clpbn_tabled_retract(:),
clpbn_tabled_abolish(:),
clpbn_tabled_asserta(:),
clpbn_tabled_assertz(:),
@ -121,6 +125,9 @@ clpbn_table(F/N,M) :-
% enter evidence after binding.
( var(A0) -> A0 = V2 ; put_evidence(A0, V2) )
;
clpbn:clpbn_flag(solver,none) ->
true
;
throw(error(tabled_clpbn_predicate_should_never_fail,S))
)
)
@ -179,7 +186,7 @@ clpbn_table_nondet(F/N,M) :-
NKey =.. [NF|Args],
asserta(clpbn_table(Key, M, NKey)),
assert(
(M:Key :- writeln(in:Key),
(M:Key :- % writeln(in:Key),
b_getval(clpbn_tables, Tab),
( b_hash_lookup(Key, Out, Tab) ->
fail
@ -217,6 +224,31 @@ clpbn_tabled_clause(Head, M, Body) :-
clpbn_table(Head, M, THead),
clause(M:THead, Body).
clpbn_tabled_clause_ref(M:Head, Body, Ref) :- !,
clpbn_tabled_clause_ref(Head, M, Body, Ref).
clpbn_tabled_clause_ref(Head, Body, Ref) :-
prolog_load_context(module, M),
clpbn_tabled_clause_ref(Head, M, Body, Ref).
clpbn_tabled_clause_ref(M:Head, _, Body, Ref) :- !,
clpbn_tabled_clause_ref(Head, M, Body, Ref).
clpbn_tabled_clause_ref(Head, M, Body, Ref) :-
clpbn_table(Head, M, THead),
clause(M:THead, Body, Ref).
clpbn_tabled_retract(M:Head) :- !,
clpbn_tabled_retract(Head, M).
clpbn_tabled_retract(Head) :-
prolog_load_context(module, M),
clpbn_tabled_retract(Head, M).
clpbn_tabled_retract(M:Head, _) :- !,
clpbn_tabled_retract(Head, M).
clpbn_tabled_retract(Head, M) :-
clpbn_table(Head, M, THead),
retract(M:THead).
clpbn_tabled_assertz(M:Clause) :- !,
clpbn_tabled_assertz2(Clause, M).

1
packages/CLPBN/clpbn/ve.yap
View File

@ -103,6 +103,7 @@ solve_ve([LVs|_], [NVs0|_], Ps) :-
sort(LV0, LV),
% construct the graph
find_all_table_deps(Tables0, LV),
%writeln((Li: LVs: LV)),
process(LVi, LVs, tab(Dist,_,_)),
%writeln(m:Dist),matrix:matrix_to_list(Dist,LD),writeln(LD),
%exps(LD,LDE),writeln(LDE),

131
packages/CLPBN/learning/aleph_params.yap
View File

@ -1,8 +1,13 @@
%
% Interface the Aleph ILP system to CLP(BN)
%
% Relies on the Aleph cost function.
% It assumes Aleph work as usual, but some variables are of type random.
% Aleph was written by Ashwin Srinivasan
%
% this code relies on hacked version of Aleph, contact
% vsc@dcc.fc.up.pt
%
% Aleph generates clauses as usual,
% but some variables are of special type random.
%
:- module(clpbn_aleph,
[init_clpbn_cost/0,
@ -33,6 +38,7 @@
clpbn_tabled_asserta/2,
clpbn_tabled_assertz/1,
clpbn_tabled_clause/2,
clpbn_tabled_clause_ref/3,
clpbn_tabled_number_of_clauses/2,
clpbn_is_tabled/1,
clpbn_reset_tables/0,
@ -42,7 +48,7 @@
% Tell Aleph not to use default solver during saturation
%
% all work will be done by EM
:- set_clpbn_flag(solver,none).
%:- set_clpbn_flag(solver,none).
%
% This is the Aleph interface
@ -58,85 +64,50 @@
:- user:set(skolem_examples, false).
% avoid doing CLP(BN) stuff except at start
:- user:set(sat_start_hook, clpbn_aleph:disable_solver).
:- user:set(sat_stop_hook, clpbn_aleph:enable_solver).
:- user:set(sat_begin_hook, clpbn_aleph:disable_solver).
:- user:set(sat_end_hook, clpbn_aleph:enable_solver).
:- user:set(reduce_start_hook, clpbn_aleph:disable_solver).
:- user:set(reduce_stop_hook, clpbn_aleph:enable_solver).
:- user:set(reduce_begin_hook, clpbn_aleph:disable_solver).
:- user:set(reduce_end_hook, clpbn_aleph:enable_solver).
:- user:set(record_testclause_hook, clpbn_aleph:do_nothing).
:- user:set(best_clause_hook, clpbn_aleph:add_new_clause).
:- user:set(newbest_hook, clpbn_aleph:store_theory).
disable_solver(_) :-
clpbn_flag(solver, Old, none),
nb_setval(old_clpbn_solver, Old).
disable_solver(_,_) :-
disable_solver.
disable_solver(_) :-
disable_solver.
disable_solver :-
clpbn_flag(solver, Old, none),
nb_setval(old_clpbn_solver, Old).
enable_solver(_) :-
enable_solver.
enable_solver(_,_) :-
enable_solver.
enable_solver :-
nb_getval(old_clpbn_solver, Old),
set_clpbn_flag(solver, Old).
enable_solver(_,_) :-
nb_getval(old_clpbn_solver, Old),
set_clpbn_flag(solver, Old).
do_nothing(_).
% backup current best theory in DB.
store_theory(_,_,_) :-
eraseall(best_theory),
% step 1: update distributions to better values.
add_new_clause(_,(H :- _),_,_) :-
(
clpbn_is_tabled(user:H)
->
update_tabled_theory(H)
;
update_theory(H)
),
fail.
store_theory(_,(H:-_),_) :-
clpbn_is_tabled(user:H), !,
store_tabled_theory(H).
store_theory(_,(H:-_),_) :- !,
store_theory(H).
store_theory(_,H,_) :-
store_theory(H).
store_tabled_theory(H) :-
clpbn_tabled_clause(user:H,B),
add_correct_cpt(B,NB),
store_cl((H:-NB)),
fail.
store_tabled_theory(_).
store_theory(H) :-
clause(user:H,B),
add_correct_cpt(B,NB),
store_cl((H:-NB)),
fail.
store_theory(_).
add_correct_cpt((G,B),(G,NB)) :- !,
add_correct_cpt(B,NB).
add_correct_cpt((clpbn:{V = K with Tab }), ({V = K with NTab})) :-
correct_tab(Tab,K,NTab).
add_correct_cpt(({V = K with Tab }), ({V = K with NTab})) :-
correct_tab(Tab,K,NTab).
correct_tab(p(Vs,_),K,p(Vs,TDist)) :-
get_dist_key(Id, K),
get_dist_params(Id, TDist).
correct_tab(p(Vs,_,Ps),K,p(Vs,TDist,Ps)) :-
get_dist_key(Id, K),
get_dist_params(Id, TDist).
store_cl(Cl) :-
recordz(best_theory, Cl, _).
:- user:set(best_clause_hook, clpbn_aleph:add_new_clause).
% step 2: add clause
add_new_clause(_,(_ :- true),_,_) :- !.
add_new_clause(_,(H :- B),_,_) :-
%trace,
% user:db_usage,
% user:db_dynamic,
domain(H, K, V, D),
user:cost((H :- B), _, Cost),
format(user_error,'Current Cost: ~w.~n', [Cost]),
user:cost((H :- B), _, _Cost),
rewrite_body(B, IB, Vs, _, ( !, { V = K with p(D, CPTList, Vs) })),
% need to remember which CPT we want
get_dist_key(Id, K),
@ -157,8 +128,38 @@ add_new_clause(_,(H :- B),_,_) :-
).
update_tabled_theory(H) :-
clpbn_tabled_clause_ref(user:H,B,Ref),
add_correct_cpt(B,NB),
erase(Ref),
clpbn_tabled_assertz((user:(H:-NB))),
fail.
update_tabled_theory(_).
update_theory(H) :-
clause(user:H,B,Ref),
add_correct_cpt(B,NB),
erase(Ref),
assert((user:H:-NB)),
fail.
update_theory(_).
add_correct_cpt((G,B),(G,NB)) :-
add_correct_cpt(B,NB).
add_correct_cpt((clpbn:{V = K with Tab }), ({V = K with NTab})) :-
correct_tab(Tab,K,NTab).
add_correct_cpt(({V = K with Tab }), ({V = K with NTab})) :-
correct_tab(Tab,K,NTab).
correct_tab(p(Vs,_),K,p(Vs,TDist)) :-
get_dist_key(Id, K),
get_dist_params(Id, TDist).
correct_tab(p(Vs,_,Ps),K,p(Vs,TDist,Ps)) :-
get_dist_key(Id, K),
get_dist_params(Id, TDist).
% user-defined cost function, Aleph knows about this (and only about this).
user:cost((H :- B),Inf,Score) :- !,
user:cost((H :- B),Inf,Score) :-
domain(H, K, V, D),
check_info(Inf),
rewrite_body(B, IB, Vs, Ds, ( !, { V = K with p(D, CPTList, Vs) })),

8
packages/CLPBN/learning/em.yap
View File

@ -5,13 +5,15 @@
:- module(clpbn_em, [em/5]).
:- use_module(library(lists),
[append/3]).
[append/3,
delete/3]).
:- use_module(library(clpbn),
[clpbn_init_graph/1,
clpbn_init_solver/5,
clpbn_run_solver/4,
clpbn_finalize_solver/1,
conditional_probability/3,
clpbn_flag/2]).
:- use_module(library('clpbn/dists'),
@ -203,8 +205,7 @@ compute_parameters([], [], _, Lik, Lik, _).
compute_parameters([Id-Samples|Dists], [Id-NewTable|Tables], MDistTable, Lik0, Lik, LPs:MargVars) :-
empty_dist(Id, Table0),
add_samples(Samples, Table0, MDistTable),
%matrix_to_list(Table0,Mat),
%format(user_error, 'FINAL ~d ~w~n', [Id,Mat]),
%matrix_to_list(Table0,Mat), lists:sumlist(Mat, Sum), format(user_error, 'FINAL ~d ~w ~w~n', [Id,Sum,Mat]),
soften_sample(Table0, SoftenedTable),
% matrix:matrix_sum(Table0,TotM),
normalise_counts(SoftenedTable, NewTable),
@ -240,3 +241,4 @@ backtrack_run_all([Item|_], Mod) :-
backtrack_run_all([_|Items], Mod) :-
backtrack_run_all(Items, Mod).
backtrack_run_all([], _).