SingleSparseInferenceBase.h

Go to the documentation of this file.

/*
 * Copyright (c) The Shogun Machine Learning Toolbox
 * Written (W) 2015 Wu Lin
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 * 1. Redistributions of source code must retain the above copyright notice, this
 *    list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright notice,
 *    this list of conditions and the following disclaimer in the documentation
 *    and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
 * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * The views and conclusions contained in the software and documentation are those
 * of the authors and should not be interpreted as representing official policies,
 * either expressed or implied, of the Shogun Development Team.
 *
 */

#ifndef CSINGLESPARSEINFERENCEBASE_H
#define CSINGLESPARSEINFERENCEBASE_H

#include <shogun/lib/config.h>

#ifdef HAVE_EIGEN3

#include <shogun/machine/gp/SparseInferenceBase.h>
#include <shogun/lib/Lock.h>

namespace shogun
{

class CSingleSparseInferenceBase: public CSparseInferenceBase
{
public:
    CSingleSparseInferenceBase();

    CSingleSparseInferenceBase(CKernel* kernel, CFeatures* features,
            CMeanFunction* mean, CLabels* labels, CLikelihoodModel* model,
            CFeatures* inducing_features);

    virtual ~CSingleSparseInferenceBase();

    virtual const char* get_name() const { return "SingleSparseInferenceBase"; }

    virtual void set_kernel(CKernel* kern);

    virtual void optimize_inducing_features();

    virtual void set_lower_bound_of_inducing_features(SGVector<float64_t> bound);

    virtual void set_upper_bound_of_inducing_features(SGVector<float64_t> bound);

    virtual void set_tolearance_for_inducing_features(float64_t tol);

    virtual void set_max_iterations_for_inducing_features(int32_t it);

    virtual void enable_optimizing_inducing_features(bool is_optmization);

protected:

    virtual float64_t get_derivative_related_cov(SGVector<float64_t> ddiagKi,
        SGMatrix<float64_t> dKuui, SGMatrix<float64_t> dKui)=0;

    virtual SGVector<float64_t> get_derivative_wrt_inducing_noise(
        const TParameter* param)=0;


    virtual SGVector<float64_t> get_derivative_wrt_inference_method(
            const TParameter* param);


    virtual SGVector<float64_t> get_derivative_wrt_kernel(
            const TParameter* param);

    virtual void check_bound(SGVector<float64_t> bound, const char* name);

    SGVector<float64_t> m_lower_bound;

    SGVector<float64_t> m_upper_bound;

    float64_t m_max_ind_iterations;

    float64_t m_ind_tolerance;

    bool m_opt_inducing_features;

    virtual void check_fully_sparse();

    virtual SGVector<float64_t> get_derivative_wrt_inducing_features(const TParameter* param)=0;

    bool m_fully_sparse;

    /* a lock used to parallelly compute derivatives wrt hyperparameters */
    CLock* m_lock;
private:
    /* init */
    void init();

    static double nlopt_function(unsigned n, const double* x, double* grad, void* func_data);
};
}
#endif /* HAVE_EIGEN3 */
#endif /* CSINGLESPARSEINFERENCEBASE_H */