Explainable RBM model architecture

Explainable-RBM.py

def rbm(movies_df):
    """
    Implement RBM architecture in TensorFlow
    :param movies_df: data frame that stores movies information
    :return: variables to be used during TensorFlow training
    """
    hiddenUnits = 100  # Number of hidden layers
    visibleUnits = len(movies_df)  # Number of visible layers

    # Create respective placeholder variables for storing visible and hidden layer biases and weights
    vb = tf.placeholder("float", [visibleUnits])  # Number of unique movies
    hb = tf.placeholder("float", [hiddenUnits])  # Number of features
    W = tf.placeholder("float", [visibleUnits, hiddenUnits])  # Weights that connect the hidden and visible layers

    # Pre-process the input data
    v0 = tf.placeholder("float", [None, visibleUnits])
    _h0 = tf.nn.sigmoid(tf.matmul(v0, W) + hb)
    h0 = tf.nn.relu(tf.sign(_h0 - tf.random_uniform(tf.shape(_h0))))

    # Reconstruct the pre-processed input data (Sigmoid and ReLU activation functions are used)
    _v1 = tf.nn.sigmoid(tf.matmul(h0, tf.transpose(W)) + vb)
    v1 = tf.nn.relu(tf.sign(_v1 - tf.random_uniform(tf.shape(_v1))))
    h1 = tf.nn.sigmoid(tf.matmul(v1, W) + hb)

    # Set RBM training parameters
    alpha = 0.1  # Set learning rate
    w_pos_grad = tf.matmul(tf.transpose(v0), h0)  # Set positive gradients
    w_neg_grad = tf.matmul(tf.transpose(v1), h1)  # Set negative gradients

    # Calculate contrastive divergence to maximize
    CD = (w_pos_grad - w_neg_grad) / tf.to_float(tf.shape(v0)[0])

    # Create methods to update the weights and biases
    update_w = W + alpha * CD
    update_vb = vb + alpha * tf.reduce_mean(v0 - v1, 0)
    update_hb = hb + alpha * tf.reduce_mean(h0 - h1, 0)

    # Set error function (RMSE)
    err = v0 - v1
    err_sum = tf.sqrt(tf.reduce_mean(err * err))

    # Initialize variables
    cur_w = np.zeros([visibleUnits, hiddenUnits], np.float32)  # Current weight
    cur_vb = np.zeros([visibleUnits], np.float32)  # Current visible unit biases
    cur_hb = np.zeros([hiddenUnits], np.float32)  # Current hidden unit biases
    prv_w = np.zeros([visibleUnits, hiddenUnits], np.float32)  # Previous weight
    prv_vb = np.zeros([visibleUnits], np.float32)  # Previous visible unit biases
    prv_hb = np.zeros([hiddenUnits], np.float32)  # Previous hidden unit biases

    return v0, W, vb, hb, update_w, prv_w, prv_vb, prv_hb, update_vb, update_hb, cur_w, cur_vb, cur_hb, err_sum