|
pulsatrix
|
The two numbers a forward pass through the bridge produces: the derived query fact's real-valued weight, and its exact partial derivative w.r.t. the neural predicate's (sigmoid-squashed) output – both computed in the same DualSemiring<double> evaluation pass (see datalog_dual_semiring.hpp).
More...
#include <neuro_symbolic_datalog_bridge.hpp>
Public Attributes | |
| double | query_weight |
| ancestor(a,d)'s derived weight. | |
| double | grad_wrt_predicate_output |
| d(query_weight) / d(edge(a,b)'s weight, i.e. the sigmoid output). | |
The two numbers a forward pass through the bridge produces: the derived query fact's real-valued weight, and its exact partial derivative w.r.t. the neural predicate's (sigmoid-squashed) output – both computed in the same DualSemiring<double> evaluation pass (see datalog_dual_semiring.hpp).
| double pulsatrix::datalog::NeuralPredicateQueryResult::grad_wrt_predicate_output |
d(query_weight) / d(edge(a,b)'s weight, i.e. the sigmoid output).
| double pulsatrix::datalog::NeuralPredicateQueryResult::query_weight |
ancestor(a,d)'s derived weight.