def test_ais_with_dbn_sanity_check(self): dbn = DBN(RBM(5, 20)) dbn.add_layer(RBM(20, 5)) dbn.add_layer(RBM(5, 20)) dbn[0].W = np.matrix(np.random.randn(5, 20)) dbn[0].b = np.matrix(np.random.rand(5, 1) - 0.5) dbn[0].c = np.matrix(np.random.rand(20, 1) - 0.5) dbn[1].W = dbn[0].W.T dbn[1].b = dbn[0].c dbn[1].c = dbn[0].b dbn[2].W = dbn[0].W dbn[2].b = dbn[0].b dbn[2].c = dbn[0].c samples = dbn.sample(100) ais = Estimator(dbn[0]) rbm_logz = ais.estimate_log_partition_function(100, np.sqrt(np.arange(0, 1, 0.001))) rbm_probs = ais.estimate_log_probability(samples) ais = Estimator(dbn) dbn_logz = ais.estimate_log_partition_function(100, np.sqrt(np.arange(0, 1, 0.001))) dbn_probs = ais.estimate_log_probability(samples) self.assertTrue(abs(dbn_logz - rbm_logz) / rbm_logz < 0.02) self.assertTrue((np.exp(dbn_probs) - np.exp(rbm_probs)).mean() < 0.02)
def test_ais_with_dbn(self): dbn = DBN(RBM(10, 10)) dbn.add_layer(RBM(10, 20)) hidden_states = utils.binary_numbers(dbn[0].Y.shape[0]) ais = Estimator(dbn) ais_logz = ais.estimate_log_partition_function(100, np.sqrt(np.arange(0, 1, 0.001))) brf_logz = utils.logsumexp(dbn[1]._ulogprob_vis(hidden_states)) brf_probs = [] dbn_probs = [] dbn_bound = [] for i in range(50): sample = np.matrix(np.random.rand(10, 1)) > 0.5 prob, bound = ais.estimate_log_probability(sample, 200) brf_probs.append(utils.logsumexp(dbn[1]._ulogprob_vis(hidden_states) + dbn[0]._clogprob_vis_hid(sample, hidden_states), 1) - brf_logz) dbn_probs.append(prob) dbn_bound.append(bound) self.assertTrue(np.mean(dbn_bound) < np.mean(dbn_probs)) self.assertTrue(np.abs(np.mean(dbn_probs) - np.mean(brf_probs)) < 0.1)
def test_ais_with_semirbm_dbn(self): dbn = DBN(RBM(5, 5)) dbn.add_layer(SemiRBM(5, 5)) ais = Estimator(dbn) ais.estimate_log_partition_function(100, np.arange(0, 1, 1E-3), layer=0) ais.estimate_log_partition_function(10, np.arange(0, 1, 1E-3), layer=1) dbn[0]._brf_logz = utils.logsumexp(dbn[0]._ulogprob_vis( utils.binary_numbers(dbn[0].X.shape[0]))) dbn[1]._brf_logz = utils.logsumexp(dbn[1]._ulogprob_vis( utils.binary_numbers(dbn[1].X.shape[0]))) samples = np.concatenate( [dbn.sample(25, 100, 20), np.matrix(np.random.rand(5, 25) > 0.5)], 1) Y = utils.binary_numbers(dbn[0].Y.shape[0]) X = utils.binary_numbers(dbn[0].X.shape[0]) logRy = dbn[1]._ulogprob_vis(Y) logQy = utils.logsumexp(dbn[0]._ulogprob(X, Y, all_pairs=True), 0) log_sum = utils.logsumexp( dbn[0]._clogprob_hid_vis(samples, Y, all_pairs=True) - logQy + logRy, 1) logPx = log_sum + dbn[0]._ulogprob_vis(samples) - dbn[1]._brf_logz logPx_ = ais.estimate_log_probability(samples)[0] self.assertTrue(np.abs(logPx_.mean() - logPx.mean()) < 0.1)
def test_ais_with_dbn(self): dbn = DBN(RBM(10, 10)) dbn.add_layer(RBM(10, 20)) hidden_states = utils.binary_numbers(dbn[0].Y.shape[0]) ais = Estimator(dbn) ais_logz = ais.estimate_log_partition_function( 100, np.sqrt(np.arange(0, 1, 0.001))) brf_logz = utils.logsumexp(dbn[1]._ulogprob_vis(hidden_states)) brf_probs = [] dbn_probs = [] dbn_bound = [] for i in range(50): sample = np.matrix(np.random.rand(10, 1)) > 0.5 prob, bound = ais.estimate_log_probability(sample, 200) brf_probs.append( utils.logsumexp( dbn[1]._ulogprob_vis(hidden_states) + dbn[0]._clogprob_vis_hid(sample, hidden_states), 1) - brf_logz) dbn_probs.append(prob) dbn_bound.append(bound) self.assertTrue(np.mean(dbn_bound) < np.mean(dbn_probs)) self.assertTrue(np.abs(np.mean(dbn_probs) - np.mean(brf_probs)) < 0.1)
def test_ais_with_dbn_sanity_check(self): dbn = DBN(RBM(5, 20)) dbn.add_layer(RBM(20, 5)) dbn.add_layer(RBM(5, 20)) dbn[0].W = np.matrix(np.random.randn(5, 20)) dbn[0].b = np.matrix(np.random.rand(5, 1) - 0.5) dbn[0].c = np.matrix(np.random.rand(20, 1) - 0.5) dbn[1].W = dbn[0].W.T dbn[1].b = dbn[0].c dbn[1].c = dbn[0].b dbn[2].W = dbn[0].W dbn[2].b = dbn[0].b dbn[2].c = dbn[0].c samples = dbn.sample(100) ais = Estimator(dbn[0]) rbm_logz = ais.estimate_log_partition_function( 100, np.sqrt(np.arange(0, 1, 0.001))) rbm_probs = ais.estimate_log_probability(samples) ais = Estimator(dbn) dbn_logz = ais.estimate_log_partition_function( 100, np.sqrt(np.arange(0, 1, 0.001))) dbn_probs = ais.estimate_log_probability(samples) self.assertTrue(abs(dbn_logz - rbm_logz) / rbm_logz < 0.02) self.assertTrue((np.exp(dbn_probs) - np.exp(rbm_probs)).mean() < 0.02)
def main(argv): num_visibles = 28 * 28 num_hiddens = [1000, 1000] num_epochs = 50 batch_size = 100 # load data samples data = load('../data/mnist.npz')['train'] / 255. # train 1st layer dbn = DBN(RBM(num_visibles, num_hiddens[0])) dbn.train(data, num_epochs, batch_size) # train 2nd layer dbn.add_layer(RBM(num_hiddens[0], num_hiddens[1])) dbn.train(data, num_epochs, batch_size, [1E-1, 1E-2]) return 0
def test_ais_with_semirbm_dbn(self): dbn = DBN(RBM(5, 5)) dbn.add_layer(SemiRBM(5, 5)) ais = Estimator(dbn) ais.estimate_log_partition_function(100, np.arange(0, 1, 1E-3), layer=0) ais.estimate_log_partition_function(10, np.arange(0, 1, 1E-3), layer=1) dbn[0]._brf_logz = utils.logsumexp(dbn[0]._ulogprob_vis(utils.binary_numbers(dbn[0].X.shape[0]))) dbn[1]._brf_logz = utils.logsumexp(dbn[1]._ulogprob_vis(utils.binary_numbers(dbn[1].X.shape[0]))) samples = np.concatenate([dbn.sample(25, 100, 20), np.matrix(np.random.rand(5, 25) > 0.5)], 1) Y = utils.binary_numbers(dbn[0].Y.shape[0]) X = utils.binary_numbers(dbn[0].X.shape[0]) logRy = dbn[1]._ulogprob_vis(Y) logQy = utils.logsumexp(dbn[0]._ulogprob(X, Y, all_pairs=True), 0) log_sum = utils.logsumexp(dbn[0]._clogprob_hid_vis(samples, Y, all_pairs=True) - logQy + logRy, 1) logPx = log_sum + dbn[0]._ulogprob_vis(samples) - dbn[1]._brf_logz logPx_ = ais.estimate_log_probability(samples)[0] self.assertTrue(np.abs(logPx_.mean() - logPx.mean()) < 0.1)
def main(argv): # load preprocessed data samples print 'loading data...\t', data = load('./data/vanhateren.npz') print '[DONE]' print # remove DC component (first component) data_train = data['train'][1:, :] data_test = data['test'][1:, :] # create 1st layer dbn = DBN(GaussianRBM(num_visibles=data_train.shape[0], num_hiddens=100)) # hyperparameters dbn[0].learning_rate = 5E-3 dbn[0].weight_decay = 1E-2 dbn[0].momentum = 0.9 dbn[0].sigma = 0.65 dbn[0].cd_steps = 1 dbn[0].persistent = True # train 1st layer print 'training...\t', dbn.train(data_train, num_epochs=100, batch_size=100) print '[DONE]' # evaluate 1st layer print 'evaluating...\t', logptf = dbn.estimate_log_partition_function(num_ais_samples=100, beta_weights=arange(0, 1, 1E-3)) loglik = dbn.estimate_log_likelihood(data_test) print '[DONE]' print print 'estimated log-partf.:\t', logptf print 'estimated log-loss:\t', -loglik / data_test.shape[0] / log(2) print # create 2nd layer dbn.add_layer(SemiRBM(num_visibles=100, num_hiddens=100)) # initialize parameters dbn[1].L = dbn[0].W.T * dbn[0].W dbn[1].b = dbn[0].W.T * dbn[0].b + dbn[0].c + 0.5 * asmatrix(diag(dbn[1].L)).T dbn[1].L = dbn[1].L - asmatrix(diag(diag(dbn[1].L))) # hyperparameters dbn[1].learning_rate = 5E-3 dbn[1].learning_rate_lateral = 5E-4 dbn[1].weight_decay = 5E-3 dbn[1].weight_decay_lateral = 5E-3 dbn[1].momentum = 0.9 dbn[1].momentum_lateral = 0.9 dbn[1].num_lateral_updates = 20 dbn[1].damping = 0.2 dbn[1].cd_steps = 1 dbn[1].persistent = True # train 2nd layer print 'training...\t', dbn.train(data_train, num_epochs=100, batch_size=100) print '[DONE]' # evaluate 2nd layer print 'evaluating...\t', logptf = dbn.estimate_log_partition_function(num_ais_samples=100, beta_weights=arange(0, 1, 1E-3)) loglik = dbn.estimate_log_likelihood(data_test, num_samples=100) print '[DONE]' print print 'estimated log-partf.:\t', logptf print 'estimated log-loss:\t', -loglik / data_test.shape[0] / log(2) print # fine-tune with wake-sleep dbn[0].learning_rate /= 4. dbn[1].learning_rate /= 4. print 'fine-tuning...\t', dbn.train_wake_sleep(data_train, num_epochs=10, batch_size=10) print '[DONE]' # reevaluate print 'evaluating...\t', logptf = dbn.estimate_log_partition_function(num_ais_samples=100, beta_weights=arange(0, 1, 1E-3)) loglik = dbn.estimate_log_likelihood(data_test, num_samples=100) print '[DONE]' print print 'estimated log-partf.:\t', logptf print 'estimated log-loss:\t', -loglik / data_test.shape[0] / log(2) return 0