NeuralLayer.h

Go to the documentation of this file.

/*
 * Copyright (c) 2014, Shogun Toolbox Foundation
 * 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.
 *
 * 3. Neither the name of the copyright holder nor the names of its
 * contributors may be used to endorse or promote products derived from this
 * software without specific prior written permission.

 * 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 HOLDER 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.
 *
 * Written (W) 2014 Khaled Nasr
 */

#ifndef __NEURALLAYER_H__
#define __NEURALLAYER_H__

#include <shogun/lib/common.h>
#include <shogun/base/SGObject.h>
#include <shogun/lib/SGMatrix.h>
#include <shogun/lib/SGVector.h>
#include <shogun/lib/DynamicObjectArray.h>

namespace shogun
{

enum ENLAutoencoderPosition
{
    NLAP_NONE=0,
    NLAP_ENCODING=1,
    NLAP_DECODING=2
};

template <class T> class SGVector;

class CNeuralLayer : public CSGObject
{
public:
    CNeuralLayer();

    CNeuralLayer(int32_t num_neurons);

    virtual ~CNeuralLayer();

    virtual void initialize_neural_layer(CDynamicObjectArray* layers,
            SGVector<int32_t> input_indices);

    virtual void set_batch_size(int32_t batch_size);

    virtual bool is_input() { return false; }

    virtual void initialize_parameters(SGVector<float64_t> parameters,
            SGVector<bool> parameter_regularizable,
            float64_t sigma) { }

    virtual void compute_activations(SGMatrix<float64_t> inputs) { }

    virtual void compute_activations(SGVector<float64_t> parameters,
            CDynamicObjectArray* layers) { }

    virtual void compute_gradients(SGVector<float64_t> parameters,
            SGMatrix<float64_t> targets,
            CDynamicObjectArray* layers,
            SGVector<float64_t> parameter_gradients) { }

    virtual float64_t compute_error(SGMatrix<float64_t> targets) { return 0; }

    virtual void enforce_max_norm(SGVector<float64_t> parameters,
            float64_t max_norm) { }

    virtual void dropout_activations();

    virtual float64_t compute_contraction_term(SGVector<float64_t> parameters)
    {
        return 0.0;
    }

    virtual int32_t get_num_neurons() { return m_num_neurons; }

    virtual int32_t get_width() { return m_width; }

    virtual int32_t get_height() { return m_height; }

    virtual void set_num_neurons(int32_t num_neurons)
    {
        m_num_neurons = num_neurons;
        set_batch_size(m_batch_size);
    }

    virtual int32_t get_num_parameters() { return m_num_parameters; }

    virtual SGMatrix<float64_t> get_activations() { return m_activations; }

    virtual SGMatrix<float64_t> get_activation_gradients()
    {
        return m_activation_gradients;
    }

    virtual SGMatrix<float64_t> get_local_gradients()
    {
        return m_local_gradients;
    }

    virtual SGVector<int32_t> get_input_indices() { return m_input_indices; }

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

private:
    void init();

public:
    bool is_training;

    float64_t dropout_prop;

    float64_t contraction_coefficient;

    ENLAutoencoderPosition autoencoder_position;

protected:
    int32_t m_num_neurons;

    int32_t m_width;

    int32_t m_height;

    int32_t m_num_parameters;

    SGVector<int32_t> m_input_indices;

    SGVector<int32_t> m_input_sizes;

    int32_t m_batch_size;

    SGMatrix<float64_t> m_activations;

    SGMatrix<float64_t> m_activation_gradients;

    SGMatrix<float64_t> m_local_gradients;

    SGMatrix<bool> m_dropout_mask;
};

}
#endif