SquaredErrorLoss.hpp

Go to the documentation of this file.

#ifndef SQUAREDLOSS_HPP_
#define SQUAREDLOSS_HPP_

#include "loss/Loss.hpp"

namespace mic {
namespace neural_nets {
namespace loss {

template <typename dtype=float>
class SquaredErrorLoss : public Loss<dtype> {
public:
    dtype calculateLoss (mic::types::MatrixPtr<dtype> target_y_, mic::types::MatrixPtr<dtype> predicted_y_) {
        // Sizes must match.
        assert(predicted_y_->size() == target_y_->size());

        // Calculate loss.
        dtype loss =0;
        for (size_t i=0; i <(size_t)predicted_y_->size(); i++) {
            loss += ((*target_y_)[i] - (*predicted_y_)[i])*((*target_y_)[i] - (*predicted_y_)[i]);
        }
        // Return squared error (SE).
        // The mean squared error (MSE) is calculated by dividing the SE by the size of a batch.
        return loss/2.0;
    }

    mic::types::MatrixPtr<dtype> calculateGradient (mic::types::MatrixPtr<dtype> target_y_, mic::types::MatrixPtr<dtype> predicted_y_) {
        // Sizes must match.
        assert(predicted_y_->size() == target_y_->size());

        // Calculate gradient.
        mic::types::MatrixPtr<dtype> dy = MAKE_MATRIX_PTR(dtype, predicted_y_->rows(), predicted_y_->cols());
        for (size_t i=0; i <(size_t)predicted_y_->size(); i++) {
            (*dy)[i] = -((*target_y_)[i] - (*predicted_y_)[i]);
        }

        /*std::cout << " predicted_y_ = " << (*predicted_y_) << std::endl;
        std::cout << " target_y_ = " << (*target_y_) << std::endl;
        std::cout << " dy = (p-t) = " << (*dy) << std::endl;*/

        return dy;
    }

};

} //: loss
} //: neural_nets
} //: mic

#endif /* SQUAREDLOSS_HPP_ */