Threshold Logic Unit (TLU)

w

1

1

w

2

1

θ

1.5

x

1

x

2

This Demonstration illustrates the concept of the simplest artificial neuron: the threshold logic unit (TLU). This pattern space represents the different possibilities that can occur whenever different inputs for the TLU are applied.

The controls

x

1

and

x

2

represent the system's inputs, while

w

1

and

w

2

are the synaptic weights. The output is evaluated with the following rule:

y

is true if and only if

w

1

x

1

+

w

2

x

2

≥θ

. The decision line is represented in blue and the weights vector is shown as a dashed line. Everything that falls below the threshold (represented by the blue area) is evaluated to a false output. The bias is represented by

θ

. The blue dots represent the ideal possibilities that the inputs can take without the presence of noise. The dot turns red whenever the output is weighted as false and green when it is true.

This Demonstration can be used to show the robustness of a basic neural system as a logic function (AND or OR) upon noisy input (different from logic 1's and 0's).

References

[1] K. Gurney, An Introduction to Neural Networks, Boca Raton, FL: CRC Press, 1997.

Permanent Citation

Hector Sanchez

"Threshold Logic Unit (TLU)"
http://demonstrations.wolfram.com/ThresholdLogicUnitTLU/
Wolfram Demonstrations Project
Published: June 17, 2011