Added sumProduct as a convenient alias
parent
f9d1af328f
commit
9eea6cf21a
|
@ -144,6 +144,23 @@ namespace gtsam {
|
||||||
boost::dynamic_pointer_cast<DiscreteConditional>(lookup), max);
|
boost::dynamic_pointer_cast<DiscreteConditional>(lookup), max);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/* ************************************************************************ */
|
||||||
|
// sumProduct is just an alias for regular eliminateSequential.
|
||||||
|
DiscreteBayesNet DiscreteFactorGraph::sumProduct(
|
||||||
|
OptionalOrderingType orderingType) const {
|
||||||
|
gttic(DiscreteFactorGraph_sumProduct);
|
||||||
|
auto bayesNet = BaseEliminateable::eliminateSequential(orderingType);
|
||||||
|
return *bayesNet;
|
||||||
|
}
|
||||||
|
|
||||||
|
DiscreteLookupDAG DiscreteFactorGraph::sumProduct(
|
||||||
|
const Ordering& ordering) const {
|
||||||
|
gttic(DiscreteFactorGraph_sumProduct);
|
||||||
|
auto bayesNet =
|
||||||
|
BaseEliminateable::eliminateSequential(ordering, EliminateForMPE);
|
||||||
|
return DiscreteLookupDAG::FromBayesNet(*bayesNet);
|
||||||
|
}
|
||||||
|
|
||||||
/* ************************************************************************ */
|
/* ************************************************************************ */
|
||||||
// The max-product solution below is a bit clunky: the elimination machinery
|
// The max-product solution below is a bit clunky: the elimination machinery
|
||||||
// does not allow for differently *typed* versions of elimination, so we
|
// does not allow for differently *typed* versions of elimination, so we
|
||||||
|
|
|
@ -132,11 +132,28 @@ class GTSAM_EXPORT DiscreteFactorGraph
|
||||||
const std::string& s = "DiscreteFactorGraph",
|
const std::string& s = "DiscreteFactorGraph",
|
||||||
const KeyFormatter& formatter = DefaultKeyFormatter) const override;
|
const KeyFormatter& formatter = DefaultKeyFormatter) const override;
|
||||||
|
|
||||||
|
/**
|
||||||
|
* @brief Implement the sum-product algorithm
|
||||||
|
*
|
||||||
|
* @param orderingType : one of COLAMD, METIS, NATURAL, CUSTOM
|
||||||
|
* @return DiscreteBayesNet encoding posterior P(X|Z)
|
||||||
|
*/
|
||||||
|
DiscreteBayesNet sumProduct(
|
||||||
|
OptionalOrderingType orderingType = boost::none) const;
|
||||||
|
|
||||||
|
/**
|
||||||
|
* @brief Implement the sum-product algorithm
|
||||||
|
*
|
||||||
|
* @param ordering
|
||||||
|
* @return DiscreteBayesNet encoding posterior P(X|Z)
|
||||||
|
*/
|
||||||
|
DiscreteLookupDAG sumProduct(const Ordering& ordering) const;
|
||||||
|
|
||||||
/**
|
/**
|
||||||
* @brief Implement the max-product algorithm
|
* @brief Implement the max-product algorithm
|
||||||
*
|
*
|
||||||
* @param orderingType : one of COLAMD, METIS, NATURAL, CUSTOM
|
* @param orderingType : one of COLAMD, METIS, NATURAL, CUSTOM
|
||||||
* @return DiscreteLookupDAG::shared_ptr DAG with lookup tables
|
* @return DiscreteLookupDAG DAG with lookup tables
|
||||||
*/
|
*/
|
||||||
DiscreteLookupDAG maxProduct(
|
DiscreteLookupDAG maxProduct(
|
||||||
OptionalOrderingType orderingType = boost::none) const;
|
OptionalOrderingType orderingType = boost::none) const;
|
||||||
|
@ -145,7 +162,7 @@ class GTSAM_EXPORT DiscreteFactorGraph
|
||||||
* @brief Implement the max-product algorithm
|
* @brief Implement the max-product algorithm
|
||||||
*
|
*
|
||||||
* @param ordering
|
* @param ordering
|
||||||
* @return DiscreteLookupDAG::shared_ptr `DAG with lookup tables
|
* @return DiscreteLookupDAG `DAG with lookup tables
|
||||||
*/
|
*/
|
||||||
DiscreteLookupDAG maxProduct(const Ordering& ordering) const;
|
DiscreteLookupDAG maxProduct(const Ordering& ordering) const;
|
||||||
|
|
||||||
|
|
|
@ -154,6 +154,16 @@ TEST(DiscreteFactorGraph, test) {
|
||||||
auto actualMPE = graph.optimize();
|
auto actualMPE = graph.optimize();
|
||||||
EXPECT(assert_equal(mpe, actualMPE));
|
EXPECT(assert_equal(mpe, actualMPE));
|
||||||
EXPECT_DOUBLES_EQUAL(9, graph(mpe), 1e-5); // regression
|
EXPECT_DOUBLES_EQUAL(9, graph(mpe), 1e-5); // regression
|
||||||
|
|
||||||
|
// Test sumProduct alias with all orderings:
|
||||||
|
auto mpeProbability = expectedBayesNet(mpe);
|
||||||
|
EXPECT_DOUBLES_EQUAL(0.28125, mpeProbability, 1e-5); // regression
|
||||||
|
for (Ordering::OrderingType orderingType :
|
||||||
|
{Ordering::COLAMD, Ordering::METIS, Ordering::NATURAL,
|
||||||
|
Ordering::CUSTOM}) {
|
||||||
|
auto bayesNet = graph.sumProduct(orderingType);
|
||||||
|
EXPECT_DOUBLES_EQUAL(mpeProbability, bayesNet(mpe), 1e-5);
|
||||||
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
/* ************************************************************************* */
|
/* ************************************************************************* */
|
||||||
|
|
Loading…
Reference in New Issue