search_branch_and_bound_partial_prediction.hpp

Go to the documentation of this file.

#ifndef FSTSEARCHBRANCHANDBOUNDPARTIALPREDICTION_H
#define FSTSEARCHBRANCHANDBOUNDPARTIALPREDICTION_H
 /* 
=========================================================================
Copyright and license:
  * FST4 software (with the exception of any bundled or externally linked
    third-party components) is Copyright (C) 2000-present Institute of
    Information Theory and Automation of the Czech Academy of Sciences (UTIA).
  * FST4 is free software: you can redistribute it and/or modify it under
    the terms of the GNU Affero General Public License as published by the
    Free Software Foundation, version 3. The full text of the license is in
    the file LICENSE in the distribution root and at
    https://www.gnu.org/licenses/agpl-3.0.html.
  * FST4 is distributed in the hope that it will be useful, but WITHOUT ANY
    WARRANTY; without even the implied warranty of MERCHANTABILITY or
    FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public
    License for more details.
  * Closed-source use without the AGPL's copyleft obligations requires a
    commercial license from UTIA; in case of interest write to
    utia@utia.cas.cz. LICENSING.md in the distribution root explains both
    options, the contributor license agreement that every contribution
    requires, and the contacts (technical questions and pull requests:
    somol@utia.cas.cz, with "FST" in the subject line).
========================================================================== */
 
#include <memory>
#include <cassert>
#include <iostream>
#include "branch_and_bound_predictor.hpp"
#include "search_branch_and_bound.hpp"
#include "error.hpp"
#include "global.hpp"
#include "result_tracker.hpp"
 
/*============== Template parameter type naming conventions ==============
--------- Numeric types: -------------------------------------------------
DATATYPE - data sample values - usually real numbers (but may be integers
          in text processing etc.)
REALTYPE - must be real numbers - for representing intermediate results of 
          calculations like mean, covariance etc.
IDXTYPE - index values for enumeration of data samples - (nonnegative) integers, 
          extent depends on numbers of samples in data
DIMTYPE - index values for enumeration of features (dimensions), or classes (not 
          class sizes) - (nonnegative) integers, usually lower extent than IDXTYPE, 
          but be aware of expressions like _classes*_features*_features ! 
          in linearized representations of feature matrices for all classes
BINTYPE - feature selection marker type - represents ca. <10 different feature 
          states (selected, deselected, sel./desel. temporarily 1st nested loop, 2nd...)
RETURNTYPE - criterion value: real value, but may be extended in future to support 
          multiple values 
--------- Class types: ---------------------------------------------------
SUBSET       - class of class type Subset 
CLASSIFIER   - class implementing interface defined in abstract class Classifier 
EVALUATOR    - class implementing interface defined in abstract class Sequential_Step 
DISTANCE     - class implementing interface defined in abstract class Distance 
DATAACCESSOR - class implementing interface defined in abstract class Data_Accessor 
INTERVALCONTAINER - class of class type Data_Intervaller 
CONTAINER    - STL container of class type Data_Interval  
========================================================================== */
 
namespace FST {

 
template<class RETURNTYPE, typename DIMTYPE, class SUBSET, class CRITERION, class PREDICTOR>
class Search_Branch_And_Bound_Partial_Prediction : public Search_Branch_And_Bound<RETURNTYPE,DIMTYPE,SUBSET,CRITERION>{
public:
        typedef Search_Branch_And_Bound<RETURNTYPE,DIMTYPE,SUBSET,CRITERION> parent; 
        typedef typename parent::PCriterion PCriterion; 
        typedef typename parent::PSubset PSubset; 
        typedef typename parent::PNode PNode; 
        typedef typename parent::Node Node; 
        typedef typename parent::NodeType NodeType; 
        typedef typename parent::PEvaluatorEngine PEvaluatorEngine; 
        explicit Search_Branch_And_Bound_Partial_Prediction(PEvaluatorEngine engine=PEvaluatorEngine()):Search_Branch_And_Bound<RETURNTYPE,DIMTYPE,SUBSET,CRITERION>(engine) {notify("Search_Branch_And_Bound_Partial_Prediction constructor.");} 
        virtual ~Search_Branch_And_Bound_Partial_Prediction() {notify("Search_Branch_And_Bound_Partial_Prediction destructor.");}
 
        // for search() implementation see parent class

        virtual std::ostream& print(std::ostream& os) const override {
                if(parent::get_engine()->is_threaded()) os << "Branch & Bound with Partial Prediction (BBPP)  [Search_Branch_And_Bound_Partial_Prediction(threads="<<parent::get_engine()->get_max_threads()<<") with " << _predictor << "]";
                else os << "Branch & Bound with Partial Prediction (BBPP)  [Search_Branch_And_Bound_Partial_Prediction() with " << _predictor << "]";
                return os;
        };
protected:
        PREDICTOR _predictor; 
 
        // the following may be overriden in descendant Branch and Bound specialization classes
        virtual void initialize(const DIMTYPE d, const DIMTYPE n, const PCriterion crit) override {_predictor.initialize(n);} 
        virtual void leaf_result(const DIMTYPE feature, const RETURNTYPE value) override; 
        virtual void pre_evaluate_availables() override;
        virtual void post_process_tree_level() override;
        virtual bool cut_possible() override;
};
 
template<class RETURNTYPE, typename DIMTYPE, class SUBSET, class CRITERION, class PREDICTOR>
void Search_Branch_And_Bound_Partial_Prediction<RETURNTYPE,DIMTYPE,SUBSET,CRITERION,PREDICTOR>::leaf_result(const DIMTYPE feature, const RETURNTYPE value)
{
        // conditionally update prediction info: the base class fires this hook after each leaf's
        // bound update, in leaf order, so the parent node's value is the right reference point
        Node &parentnode=parent::get_parent_node();
        if(parentnode.type==parent::COMPUTED) _predictor.learn(feature,parentnode.value-value);
}
 
template<class RETURNTYPE, typename DIMTYPE, class SUBSET, class CRITERION, class PREDICTOR>
void Search_Branch_And_Bound_Partial_Prediction<RETURNTYPE,DIMTYPE,SUBSET,CRITERION,PREDICTOR>::pre_evaluate_availables()
{
        assert(parent::get_criterion());
        assert(parent::get_engine());
        const PSubset &currentset=parent::get_currentset(); assert(currentset);
        Node &parentnode=parent::get_parent_node(); assert(parentnode.type==parent::COMPUTED);
 
        // pass 1 (master thread): decide predict-vs-compute for every available feature. Deciding
        // for all of them upfront gives the same decisions as deciding one at a time while
        // evaluating, because predictor state is per-feature (the PREDICTOR contract) and each
        // feature is decided once - so the outcome does not depend on the evaluation engine
        PNode avail;
        RETURNTYPE val;
        DIMTYPE to_compute=0;
        for(bool got=this->getFirstAvailable(avail);got;got=this->getNextAvailable(avail))
        {
                if(_predictor.predict(avail->feature,val)) { //value could be predicted
                        avail->value=parentnode.value-val;
                        avail->type=parent::PREDICTED;
                } else { //value could not be predicted
                        avail->type=parent::UNKNOWN; // computed by the engine batch below
                        ++to_compute;
                }
        }
        if(to_compute==0) return;
 
        // pass 2: batch-evaluate the non-predictable features through the engine; the per-result
        // callback fires in add-order under both engine implementations, so node values are
        // assigned and prediction contrasts learned in the same order either way
        typename parent::EVALUATORENGINE::ResultCallback on_result=[this,&parentnode](const RETURNTYPE value, const PSubset&, const DIMTYPE feat) {
                PNode node;
                if(this->getAvailable(feat,node)) {
                        node->value=value;
                        node->type=parent::COMPUTED;
                        // conditionally update prediction info
                        if(parentnode.type==parent::COMPUTED) _predictor.learn(feat,parentnode.value-value);
                }
        };
        parent::get_engine()->initialize(parent::get_criterion(),parent::get_n(),on_result);
        for(bool got=this->getFirstAvailable(avail);got;got=this->getNextAvailable(avail))
        {
                if(avail->type!=parent::PREDICTED) {
                        assert(currentset->selected_raw(avail->feature));
                        currentset->deselect_raw(avail->feature);
                        if(!parent::get_engine()->add_candidate(currentset,avail->feature)) throw FST::fst_error("Criterion evaluation failure.");
                        currentset->select_raw(avail->feature);
                }
        }
        // collect() is where a worker's exception is rethrown; a false return means some
        // evaluate() failed and no callbacks fired at all, which would leave nodes at UNKNOWN and
        // poison the level sort-in - so it must be an error, not a silent continuation
        if(!parent::get_engine()->collect()) throw FST::fst_error("Criterion evaluation failure.");
}
 
template<class RETURNTYPE, typename DIMTYPE, class SUBSET, class CRITERION, class PREDICTOR>
void Search_Branch_And_Bound_Partial_Prediction<RETURNTYPE,DIMTYPE,SUBSET,CRITERION,PREDICTOR>::post_process_tree_level()
{
        assert(parent::get_criterion());
        assert(parent::get_engine());
        const PSubset &currentset=parent::get_currentset();     assert(currentset);
        Node &parentnode=parent::get_parent_node();     assert(parentnode.type==parent::PREDICTED || parentnode.type==parent::COMPUTED);
 
        // supply the true criterion value of the level's chosen PREDICTED nodes (partial
        // prediction: predictions only order the level, the tree path is exact) - batched
        // through the engine. Results map back to their nodes by submission order: the
        // per-result callback fires in add-order under both engine implementations, and the
        // submitted vector is fully built before the first add_candidate, which matters because
        // the sequential engine fires callbacks synchronously inside add_candidate. The list
        // node pointers stay valid throughout (nothing is inserted into or erased from the level
        // during the batch), and prediction contrasts are learned in the same order either way
        PNode nod;
        std::vector<PNode> submitted;
        for(bool got=this->getFirstNode(nod);got;got=this->getNextNode(nod))
                if(nod->type!=parent::COMPUTED) submitted.push_back(nod);
        if(submitted.empty()) return;
 
        DIMTYPE done=0;
        typename parent::EVALUATORENGINE::ResultCallback on_result=[this,&parentnode,&submitted,&done](const RETURNTYPE value, const PSubset&, const DIMTYPE feat) {
                PNode node=submitted[done++];
                assert(node->feature==feat);
                node->value=value;
                node->type=parent::COMPUTED;
                // conditionally update prediction info
                if(parentnode.type==parent::COMPUTED) _predictor.learn(feat,parentnode.value-value);
        };
        parent::get_engine()->initialize(parent::get_criterion(),(DIMTYPE)submitted.size(),on_result);
        for(PNode snod : submitted)
        {
                assert(currentset->selected_raw(snod->feature));
                currentset->deselect_raw(snod->feature);
                if(!parent::get_engine()->add_candidate(currentset,snod->feature)) throw FST::fst_error("Criterion evaluation failure.");
                currentset->select_raw(snod->feature);
        }
        // a failed batch must not be allowed to leave nodes without a true value on the tree path
        if(!parent::get_engine()->collect()) throw FST::fst_error("Criterion evaluation failure.");
}
 
template<class RETURNTYPE, typename DIMTYPE, class SUBSET, class CRITERION, class PREDICTOR>
bool Search_Branch_And_Bound_Partial_Prediction<RETURNTYPE,DIMTYPE,SUBSET,CRITERION,PREDICTOR>::cut_possible()
{
        if(parent::is_bound_valid()) {
                Node &currentnode=parent::get_current_node(); assert(currentnode.type==parent::COMPUTED);
                if(currentnode.value<=parent::get_bound_value()) { // BBPP ensures that the value must have been COMPUTED by now
                        return true;
                }
        }
        return false;
}
 
} // namespace
#endif // FSTSEARCHBRANCHANDBOUNDPARTIALPREDICTION_H ///:~