Trains a Siamese MLP on pairs of digits from the MNIST dataset.

It follows Hadsell-et-al.'06 [1] by computing the Euclidean distance on the output of the shared network and by optimizing the contrastive loss (see paper for more details).

# References

• Dimensionality Reduction by Learning an Invariant Mapping http://yann.lecun.com/exdb/publis/pdf/hadsell-chopra-lecun-06.pdf

Gets to 97.2% test accuracy after 20 epochs. 2 seconds per epoch on a Titan X Maxwell GPU

``````from __future__ import absolute_import
from __future__ import print_function
import numpy as np

import random
from keras.datasets import mnist
from keras.models import Model
from keras.layers import Input, Flatten, Dense, Dropout, Lambda
from keras.optimizers import RMSprop
from keras import backend as K

num_classes = 10
epochs = 20

def euclidean_distance(vects):
x, y = vects
sum_square = K.sum(K.square(x - y), axis=1, keepdims=True)
return K.sqrt(K.maximum(sum_square, K.epsilon()))

def eucl_dist_output_shape(shapes):
shape1, shape2 = shapes
return (shape1[0], 1)

def contrastive_loss(y_true, y_pred):
'''
margin = 1
square_pred = K.square(y_pred)
margin_square = K.square(K.maximum(margin - y_pred, 0))
return K.mean(y_true * square_pred + (1 - y_true) * margin_square)

def create_pairs(x, digit_indices):
'''Positive and negative pair creation.
Alternates between positive and negative pairs.
'''
pairs = []
labels = []
n = min([len(digit_indices[d]) for d in range(num_classes)]) - 1
for d in range(num_classes):
for i in range(n):
z1, z2 = digit_indices[d][i], digit_indices[d][i + 1]
pairs += [[x[z1], x[z2]]]
inc = random.randrange(1, num_classes)
dn = (d + inc) % num_classes
z1, z2 = digit_indices[d][i], digit_indices[dn][i]
pairs += [[x[z1], x[z2]]]
labels += [1, 0]
return np.array(pairs), np.array(labels)

def create_base_network(input_shape):
'''Base network to be shared (eq. to feature extraction).
'''
input = Input(shape=input_shape)
x = Flatten()(input)
x = Dense(128, activation='relu')(x)
x = Dropout(0.1)(x)
x = Dense(128, activation='relu')(x)
x = Dropout(0.1)(x)
x = Dense(128, activation='relu')(x)
return Model(input, x)

def compute_accuracy(y_true, y_pred):
'''Compute classification accuracy with a fixed threshold on distances.
'''
pred = y_pred.ravel() < 0.5
return np.mean(pred == y_true)

def accuracy(y_true, y_pred):
'''Compute classification accuracy with a fixed threshold on distances.
'''
return K.mean(K.equal(y_true, K.cast(y_pred < 0.5, y_true.dtype)))

# the data, split between train and test sets
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
input_shape = x_train.shape[1:]

# create training+test positive and negative pairs
digit_indices = [np.where(y_train == i)[0] for i in range(num_classes)]
tr_pairs, tr_y = create_pairs(x_train, digit_indices)

digit_indices = [np.where(y_test == i)[0] for i in range(num_classes)]
te_pairs, te_y = create_pairs(x_test, digit_indices)

# network definition
base_network = create_base_network(input_shape)

input_a = Input(shape=input_shape)
input_b = Input(shape=input_shape)

# because we re-use the same instance `base_network`,
# the weights of the network
# will be shared across the two branches
processed_a = base_network(input_a)
processed_b = base_network(input_b)

distance = Lambda(euclidean_distance,
output_shape=eucl_dist_output_shape)([processed_a, processed_b])

model = Model([input_a, input_b], distance)

# train
rms = RMSprop()
model.compile(loss=contrastive_loss, optimizer=rms, metrics=[accuracy])
model.fit([tr_pairs[:, 0], tr_pairs[:, 1]], tr_y,
batch_size=128,
epochs=epochs,
validation_data=([te_pairs[:, 0], te_pairs[:, 1]], te_y))

# compute final accuracy on training and test sets
y_pred = model.predict([tr_pairs[:, 0], tr_pairs[:, 1]])
tr_acc = compute_accuracy(tr_y, y_pred)
y_pred = model.predict([te_pairs[:, 0], te_pairs[:, 1]])
te_acc = compute_accuracy(te_y, y_pred)

print('* Accuracy on training set: %0.2f%%' % (100 * tr_acc))
print('* Accuracy on test set: %0.2f%%' % (100 * te_acc))
``````