Deep Learning by means of Design Pattern

pydbm is Python3 library for building restricted boltzmann machine, deep boltzmann machine, and multi-layer neural networks.

Description

The function of this library is building and modeling restricted boltzmann machine, deep boltzmann machine, and multi-layer neural networks. The models are functionally equivalent to stacked auto-encoder. The main function is the same as dimensions reduction(or pre-training).

Design thought

In relation to my Automatic Summarization Library, it is important for me that the models are functionally equivalent to stacked auto-encoder. The main function I observe is the same as dimensions reduction(or pre-training). But the functional reusability of the models can be not limited to this. These Python Scripts can be considered a kind of experiment result to verify effectiveness of object-oriented analysis, object-oriented design, and GoF's design pattern in designing and modeling neural network, deep learning, and reinforcement-Learning.

For instance, demo_dbm_multi_layer_builder.py is implemented for running the deep boltzmann machine to extract so-called feature points. This script is premised on a kind of builder pattern for separating the construction of complex restricted boltzmann machines from its graph representation so that the same construction process can create different representations. Because of common design pattern and polymorphism, the stacked auto-encoder in demo_stacked_auto_encoder.py and the multi-layer neural networks in demo_nn_multi_layer_builder.py are functionally equivalent to deep boltzmann machine.

Documentation

Full documentation is available on https://code.accel-brain.com/Deep-Learning-by-means-of-Design-Pattern/ . This document contains information on functionally reusability, functional scalability and functional extensibility.

Installation

Install using pip:

pip install pydbm

Or, you can install from wheel file.

pip install https://storage.googleapis.com/accel-brain-code/Deep-Learning-by-means-of-Design-Pattern/pydbm-1.0.6-cp36-cp36m-linux_x86_64.whl

Source code

The source code is currently hosted on GitHub.

Python package index(PyPI)

Installers for the latest released version are available at the Python package index.

Dependencies

  • numpy: v1.13.3 or higher.
  • cython: v0.27.1 or higher.
  • multipledispatch: v0.4.9

Usecase: Building the deep boltzmann machine for feature extracting.

Import Python and Cython modules.

# The `Client` in Builder Pattern
from pydbm.dbm.deep_boltzmann_machine import DeepBoltzmannMachine
# The `Concrete Builder` in Builder Pattern.
from pydbm.dbm.builders.dbm_multi_layer_builder import DBMMultiLayerBuilder
# Contrastive Divergence for function approximation.
from pydbm.approximation.contrastive_divergence import ContrastiveDivergence
# Logistic Function as activation function.
from pydbm.activation.logistic_function import LogisticFunction

instantiate objects and call the method.

dbm = DeepBoltzmannMachine(
    DBMMultiLayerBuilder(),
    # Dimention in visible layer, hidden layer, and second hidden layer.
    [traning_x.shape[1], 10, traning_x.shape[1]],
    LogisticFunction(), # Setting object for activation function.
    ContrastiveDivergence(), # Setting the object for function approximation.
    0.05 # Setting learning rate.
)
# Execute learning.
dbm.learn(traning_arr, traning_count=1000)

And the feature points can be extracted by this method.

print(dbm.get_feature_point_list(0))

Usecase: Extracting all feature points for dimensions reduction(or pre-training)

Import Python and Cython modules.

# `StackedAutoEncoder` is-a `DeepBoltzmannMachine`.
from pydbm.dbm.deepboltzmannmachine.stacked_auto_encoder import StackedAutoEncoder
# The `Concrete Builder` in Builder Pattern.
from pydbm.dbm.builders.dbm_multi_layer_builder import DBMMultiLayerBuilder
# Contrastive Divergence for function approximation.
from pydbm.approximation.contrastive_divergence import ContrastiveDivergence
# Logistic Function as activation function.
from pydbm.activation.logistic_function import LogisticFunction

instantiate objects and call the method.

dbm = StackedAutoEncoder(
    DBMMultiLayerBuilder(),
    # Dimention in visible layer, hidden layer, and second hidden layer.
    [traning_x.shape[1], 10, traning_x.shape[1]],
    LogisticFunction(), # Setting object for activation function.
    ContrastiveDivergence(), # Setting the object for function approximation.
    0.05 # Setting learning rate.
)
# Execute learning.
dbm.learn(traning_arr, traning_count=1000)

And the result of dimention reduction can be extracted by this property.

pre_trained_arr = dbm.feature_points_arr

More detail demos

Version

  • 1.0.6

Author

  • chimera0(RUM)

Author URI

  • http://accel-brain.com/

License

  • GNU General Public License v2.0