Adaptron

The Adaptron is an artificial neuron or learning device akin to the Perceptron.

Whereas the Perceptron learns the relationship between a set of independent and dependent variables by successive adjustment of weights through a process of looping:

i. Multiplying the independent variable values by the weights,
ii. Summing the products,
iii. Comparing the summation with a threshold value,
iv. Determine a calculated output,
v. Multiplying the difference between the desired and calculated output values,
vi. Learning rate and adding the results to the previous weight values,
vii. Obtaining new weight values for repetition of the process with new input values,

the Adaptron learns the precise relationship between the values of the dependent variable and the values of the independent variables by using the values of the independent variable to address the location where the dependent variable will be written to memory as data and stored in a single epoch of the data.

Once the epoch is complete the addresses may be scanned for the dependent variable data or the dependent variable data may be extracted and otherwise stored or used as a rule to replicate the relationship between the dependent and independent variables.

Thus sensor values (independent variable values) are not treated or stored as data but rather treated and stored as addresses of the corresponding dependent variable value which is stored as data at the location to which the sensor data points.

Enhanced capability may be added such as the sequence of the input variables.

A practical application example may be found here.

Advertisements

Order 7/7 & 24/24 support from here:

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s