LocalTangentSpaceAlignment.cpp

Go to the documentation of this file.

/*
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 3 of the License, or
 * (at your option) any later version.
 *
 * Written (W) 2011 Sergey Lisitsyn
 * Copyright (C) 2011 Berlin Institute of Technology and Max-Planck-Society
 */

#include <shogun/converter/LocalTangentSpaceAlignment.h>
#ifdef HAVE_LAPACK
#include <shogun/mathematics/lapack.h>
#include <shogun/lib/Time.h>
#include <shogun/lib/common.h>
#include <shogun/mathematics/Math.h>
#include <shogun/io/SGIO.h>
#include <shogun/distance/Distance.h>
#include <shogun/lib/Signal.h>

#ifdef HAVE_PTHREAD
#include <pthread.h>
#endif

using namespace shogun;

#ifndef DOXYGEN_SHOULD_SKIP_THIS
struct LTSA_THREAD_PARAM
{
    int32_t idx_start;
    int32_t idx_step;
    int32_t idx_stop;
    int32_t m_k;
    int32_t target_dim;
    const int32_t* neighborhood_matrix;
    float64_t* G_matrix;
    float64_t* mean_vector;
    float64_t* local_feature_matrix;
    const float64_t* feature_matrix;
    float64_t* s_values_vector;
    float64_t* q_matrix;
    float64_t* W_matrix;
    int32_t N;
    int32_t dim;
    int32_t actual_k;
#ifdef HAVE_PTHREAD

    PTHREAD_LOCK_T* W_matrix_lock;
#endif
};
#endif

CLocalTangentSpaceAlignment::CLocalTangentSpaceAlignment() :
        CLocallyLinearEmbedding()
{
}

CLocalTangentSpaceAlignment::~CLocalTangentSpaceAlignment()
{
}

const char* CLocalTangentSpaceAlignment::get_name() const
{
    return "LocalTangentSpaceAlignment";
};

SGMatrix<float64_t> CLocalTangentSpaceAlignment::construct_weight_matrix(CDenseFeatures<float64_t>* simple_features, float64_t* W_matrix,
                                                                         SGMatrix<int32_t> neighborhood_matrix)
{
    int32_t N = simple_features->get_num_vectors();
    int32_t dim = simple_features->get_num_features();
    int32_t t;
#ifdef HAVE_PTHREAD
    int32_t num_threads = parallel->get_num_threads();
    ASSERT(num_threads>0);
    // allocate threads and params
    pthread_t* threads = SG_MALLOC(pthread_t, num_threads);
    LTSA_THREAD_PARAM* parameters = SG_MALLOC(LTSA_THREAD_PARAM, num_threads);
#else
    int32_t num_threads = 1;
#endif

    // init matrices and norm factor to be used
    float64_t* local_feature_matrix = SG_MALLOC(float64_t, m_k*dim*num_threads);
    float64_t* mean_vector = SG_MALLOC(float64_t, dim*num_threads);
    float64_t* q_matrix = SG_MALLOC(float64_t, m_k*m_k*num_threads);
    float64_t* s_values_vector = SG_MALLOC(float64_t, dim*num_threads);
    float64_t* G_matrix = SG_MALLOC(float64_t, m_k*(1+m_target_dim)*num_threads);

    // get feature matrix
    SGMatrix<float64_t> feature_matrix = simple_features->get_feature_matrix();

#ifdef HAVE_PTHREAD
    PTHREAD_LOCK_T W_matrix_lock;
    pthread_attr_t attr;
    PTHREAD_LOCK_INIT(&W_matrix_lock);
    pthread_attr_init(&attr);
    pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_JOINABLE);

    for (t=0; t<num_threads; t++)
    {
        parameters[t].idx_start = t;
        parameters[t].idx_step = num_threads;
        parameters[t].idx_stop = N;
        parameters[t].m_k = m_k;
        parameters[t].target_dim = m_target_dim;
        parameters[t].neighborhood_matrix = neighborhood_matrix.matrix;
        parameters[t].G_matrix = G_matrix + (m_k*(1+m_target_dim))*t;
        parameters[t].mean_vector = mean_vector + dim*t;
        parameters[t].local_feature_matrix = local_feature_matrix + (m_k*dim)*t;
        parameters[t].feature_matrix = feature_matrix.matrix;
        parameters[t].s_values_vector = s_values_vector + dim*t;
        parameters[t].q_matrix = q_matrix + (m_k*m_k)*t;
        parameters[t].W_matrix = W_matrix;
        parameters[t].W_matrix_lock = &W_matrix_lock;
        parameters[t].N = N;
        parameters[t].dim = dim;
        parameters[t].actual_k = neighborhood_matrix.num_rows;
        pthread_create(&threads[t], &attr, run_ltsa_thread, (void*)&parameters[t]);
    }
    for (t=0; t<num_threads; t++)
        pthread_join(threads[t], NULL);
    PTHREAD_LOCK_DESTROY(&W_matrix_lock);
    SG_FREE(parameters);
    SG_FREE(threads);
#else
    LTSA_THREAD_PARAM single_thread_param;
    single_thread_param.idx_start = 0;
    single_thread_param.idx_step = 1;
    single_thread_param.idx_stop = N;
    single_thread_param.m_k = m_k;
    single_thread_param.target_dim = m_target_dim;
    single_thread_param.neighborhood_matrix = neighborhood_matrix.matrix;
    single_thread_param.G_matrix = G_matrix;
    single_thread_param.mean_vector = mean_vector;
    single_thread_param.local_feature_matrix = local_feature_matrix;
    single_thread_param.feature_matrix = feature_matrix;
    single_thread_param.s_values_vector = s_values_vector;
    single_thread_param.q_matrix = q_matrix;
    single_thread_param.W_matrix = W_matrix;
    single_thread_param.N = N;
    single_thread_param.dim = dim;
    single_thread_param.actual_k = neighborhood_matrix.num_rows;
    run_ltsa_thread((void*)&single_thread_param);
#endif

    // clean
    SG_FREE(G_matrix);
    SG_FREE(s_values_vector);
    SG_FREE(mean_vector);
    SG_FREE(local_feature_matrix);
    SG_FREE(q_matrix);

    int32_t actual_k = neighborhood_matrix.num_rows;
    for (int32_t i=0; i<N; i++)
    {
        for (int32_t j=0; j<m_k; j++)
            W_matrix[N*neighborhood_matrix[i*actual_k+j]+neighborhood_matrix[i*actual_k+j]] += 1.0;
    }

    return SGMatrix<float64_t>(W_matrix,N,N);
}

void* CLocalTangentSpaceAlignment::run_ltsa_thread(void* p)
{
    LTSA_THREAD_PARAM* parameters = (LTSA_THREAD_PARAM*)p;
    int32_t idx_start = parameters->idx_start;
    int32_t idx_step = parameters->idx_step;
    int32_t idx_stop = parameters->idx_stop;
    int32_t m_k = parameters->m_k;
    int32_t target_dim = parameters->target_dim;
    const int32_t* neighborhood_matrix = parameters->neighborhood_matrix;
    float64_t* G_matrix = parameters->G_matrix;
    float64_t* mean_vector = parameters->mean_vector;
    float64_t* local_feature_matrix = parameters->local_feature_matrix;
    const float64_t* feature_matrix = parameters->feature_matrix;
    float64_t* s_values_vector = parameters->s_values_vector;
    float64_t* q_matrix = parameters->q_matrix;
    float64_t* W_matrix = parameters->W_matrix;
#ifdef HAVE_PTHREAD
    PTHREAD_LOCK_T* W_matrix_lock = parameters->W_matrix_lock;
#endif

    int32_t i,j,k;
    int32_t N = parameters->N;
    int32_t dim = parameters->dim;
    int32_t actual_k = parameters->actual_k;

    for (j=0; j<m_k; j++)
        G_matrix[j] = 1.0/CMath::sqrt((float64_t)m_k);

    for (i=idx_start; i<idx_stop; i+=idx_step)
    {
        // fill mean vector with zeros
        memset(mean_vector,0,sizeof(float64_t)*dim);

        // compute local feature matrix containing neighbors of i-th vector
        for (j=0; j<m_k; j++)
        {
            for (k=0; k<dim; k++)
                local_feature_matrix[j*dim+k] = feature_matrix[neighborhood_matrix[i*actual_k+j]*dim+k];

            cblas_daxpy(dim,1.0,local_feature_matrix+j*dim,1,mean_vector,1);
        }

        // compute mean
        cblas_dscal(dim,1.0/m_k,mean_vector,1);

        // center feature vectors by mean
        for (j=0; j<m_k; j++)
            cblas_daxpy(dim,-1.0,mean_vector,1,local_feature_matrix+j*dim,1);

        int32_t info = 0;
        // find right eigenvectors of local_feature_matrix
        wrap_dgesvd('N','O',dim,m_k,local_feature_matrix,dim,
                    s_values_vector,NULL,1, NULL,1,&info);
        ASSERT(info==0);

        for (j=0; j<target_dim; j++)
        {
            for (k=0; k<m_k; k++)
                G_matrix[(j+1)*m_k+k] = local_feature_matrix[k*dim+j];
        }

        // compute GG'
        cblas_dgemm(CblasColMajor,CblasNoTrans,CblasTrans,
                    m_k,m_k,1+target_dim,
                    1.0,G_matrix,m_k,
                        G_matrix,m_k,
                    0.0,q_matrix,m_k);

        // W[neighbors of i, neighbors of i] = I - GG'
#ifdef HAVE_PTHREAD
        PTHREAD_LOCK(W_matrix_lock);
#endif
        for (j=0; j<m_k; j++)
        {
            for (k=0; k<m_k; k++)
                W_matrix[N*neighborhood_matrix[i*actual_k+k]+neighborhood_matrix[i*actual_k+j]] -= q_matrix[j*m_k+k];
        }
#ifdef HAVE_PTHREAD
        PTHREAD_UNLOCK(W_matrix_lock);
#endif
    }
    return NULL;
}

#endif /* HAVE_LAPACK */