Popular recipes tagged "neural"

Teach your computer a few tricks (Python)

2014-09-11T06:25:14-07:00

Python recipe 578932 by Alexander Pletzer (anl, back, network, neural, propagation).

Following is an artifical neural network program that takes any number of inputs and any number of hidden layers, and spits out an output. It applies back propagation with regularization to minimize the cost function. A gradient descent algorithm tries to find the minimum of the cost function in the landscape of weights.

Simple Back-propagation Neural Network in Python source code (Python)

2012-05-30T17:09:49-07:00

Python recipe 578148 by David Adler (back, back_propagation, neural, neural_network, propagation, python).

This is a slightly different version of this http://arctrix.com/nas/python/bpnn.py

Genetic Algorithm Neural Network in Python Source Code (Python)

2012-08-16T16:31:12-07:00

Python recipe 578241 by David Adler (artificial_intelligence, neural, neural_network).

A simple genetic algorithm neural network.

Artificial Neuroglial Network (ANGN) (Python)

2012-10-02T16:18:36-07:00

Python recipe 578242 by David Adler (artificial_intelligence, genetic_algorithm, genetic_algorithms, neural, neural_networks). Revision 5.

This is an attempt at emulating the algorithm from these scientific articles:

The objective of the program is to train a neural network to classify the four inputs (the dimensions of a flower) into one of three categories (three species of flower), (taken from the Iris Data Set from the UCI Machine Learning Repository). This program has two learning phases: the first is a genetic algorithm (supervised), the second is a neuroglial algorithm (unsupervised). This ANGN is a development of a previous program only consisting of a genetic algorithm which can be found here.

The second phase aims to emulate astrocytic interaction with neurons in the brain. The algorithm is based on two axioms: a) astrocytes are activated by persistent neuronal activity b) astrocytic effects occur over a longer time-scale than neurons. Each neuron has an associated astrocyte which counts the number of times its associated neuron fires (+1 for active -1 for inactive). If the counter reaches its threshold (defined as Athresh) the astrocyte is activated and for the next x iterations (defined as Adur) the astrocyte modifies the incoming weights to that particular neuron. If the counter reached a maximum due to persistent firing the incoming weights are increase by 25% for the proceeding Adur iterations; conversely if the counter reached a minimum due to persistent lack of firing the weights are decreased by 50% for the following Adur iterations). For a detailed description of the algorithm see the linked articles. For a general understanding of how this program was coded look at the pseudo-code/schematic here.

Any comments for improvements are welcome. There are several issues in this program which require addressing, please scroll down below code to read about these issues.

Genetic Algorithm in Python source code - AI-Junkie tutorial (Python)

2012-06-19T12:59:13-07:00

Python recipe 578128 by David Adler (algorithm, artificial, genetic, network, neural, python). Revision 5.

A simple genetic algorithm program. I followed this tutorial to make the program http://www.ai-junkie.com/ga/intro/gat1.html.

The objective of the code is to evolve a mathematical expression which calculates a user-defined target integer.

KEY:

chromosome = binary list (this is translated/decoded into a protein in the format number --> operator --> number etc, any genes (chromosome is read in blocks of four) which do not conform to this are ignored.

protein = mathematical expression (this is evaluated from left to right in number + operator blocks of two)

output = output of protein (mathematical expression)

error = inverse of difference between output and target

fitness score = a fraction of sum of of total errors

OTHER:

One-point crossover is used.

I have incorporated elitism in my code, which somewhat deviates from the tutorial but made my code more efficient (top ~7% of population are carried through to next generation)