Artificial Neuron

• US 20090132452A1
• Filed: 02/05/2008
• Published: 05/21/2009
• Est. Priority Date: 02/05/2007
• Status: Active Grant

• Current Assignee: Steve Adkins
• Original Assignee: Steve Adkins
• Inventors: Adkins, Steve
• Granted Patent: US 7,672,918 B2
• Time in Patent Office: Days
• US Class Current: 706/25
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs

Abstract

Artificial neurons and processing elements for artificial neurons are disclosed. One processing element generates a continuous value signal based on the first plurality of inputs and generates a responsiveness based on a second plurality of inputs. An output value determining portion generates an output signal that is equal to a predetermined value when the responsiveness signal corresponds to a non-responsive and equal to the continuous value signal when the responsiveness signal corresponds to a responsive state. Another processing element produces an output signal having a magnitude equal to zero except during a fixed time after an event when the output signal has a magnitude based on an event time.

Claims

1. A processing element for an artificial neuron, comprising:

a first plurality of inputs;

a continuous value determining portion configured to generate a continuous value signal based on the first plurality of inputs;

a second plurality of inputs;

a responsiveness determining portion configured to generate a responsiveness signal corresponding to either a responsive state or a non-responsive state based on the second plurality of inputs; and

an output value determining portion configured to receive the continuous value signal and the responsiveness signal and generate an output signal, wherein the output signal is equal to a predetermined value when the responsiveness signal corresponds to the non-responsive state and the output signal is equal to the continuous value signal when the responsiveness signal corresponds to the responsive state.

2. The processing element of claim 1, wherein the continuous value signal is equal to the multiplicative product of a first pair of inputs of the first plurality of inputs subtracted from the multiplicative product of a second pair of inputs of the first plurality of inputs.

3. The processing element of claim 1, wherein the responsiveness determining portion generates a threshold value based on a pair of inputs of the second plurality of inputs, wherein the responsiveness signal corresponds to the non-responsive state when a first input of the second plurality of inputs is less than the threshold value and the responsiveness signal corresponds to the responsive state when a second input of the second plurality of inputs is less than the threshold value.

4. The processing element of claim 3, wherein the state of the responsiveness signal does not change if the first input of the second plurality of inputs is greater than the threshold value and the second input of the second plurality of inputs is greater than the threshold value.

5. The processing element of claim 4, wherein the threshold value is based on a third input of the second plurality of inputs and a fourth input of the second plurality of inputs.

6. The processing element of claim 5, wherein the threshold value is equal to the third input subtracted from the fourth input.

7. The processing element of claim 1, wherein the first plurality of inputs and the second plurality of inputs have values corresponding to the concentrations of respective solutions of a plurality of solutions.

8. The processing element of claim 7, further comprising:

an actuator configured to receive the output signal and change the concentration of an output solution based on the output signal.

9. A processing element for an artificial neuron, comprising:

a first portion that receives a first input signal, the first portion having an activating state and a non-activating state;

a second portion that receives a second input signal, the second portion having an activating state and a non-activating state, wherein a first event is defined as the second input entering the activating state from the non-activating state while the first input is in the activating state, and a second event is defined as the first input entering the non-activating state from the activating state while the second input is in the activating state; and

a third portion that produces an output signal, the third portion having a first state where the magnitude of the output signal is equal to zero and a second state where the magnitude of the output signal and a second state where the magnitude of the output is based on an event time equal to the time elapsed between the first event and the second event.

10. The processing element of claim 9, wherein the third portion enters the second state at the time of the second event and returns to the first state after a predetermined time period has elapsed.

11. The processing element of claim 9, wherein the magnitude of the output is based on the difference of the event time and a predetermined value.

12. The processing element of claim 9, wherein the magnitude of the output signal is equal to zero if the event time is greater than a predetermined value.

13. The processing element of claim 12, wherein the magnitude of the output signal is proportional to the absolute value of the difference of the event time and a predetermined value if the event time is less than the predetermined value.

14. The processing element of claim 12, wherein the magnitude of the output signal is monotonically related to the absolute value of the difference of the event time and a predetermined value if the event time is less than the predetermined value.

15. The processing element of claim 9, wherein the magnitude of the output signal is equal to zero if the event time is less than a predetermined value.

16. The processing element of claim 15, wherein the magnitude of the output signal is proportional to the absolute value of the difference of the event time and a predetermined value if the event time is greater than the predetermined value.

17. The processing element of claim 15, wherein the magnitude of the output signal is monotonically related to the absolute value of the difference of the event time and a predetermined value if the event time is greater than the predetermined value.

18. A processing method, comprising:

receiving a first signal;

receiving a second signal;

measuring as an event time the time elapsed between a first event, wherein the second signal exceeds a second threshold value while the first signal exceeds a first threshold value, and a second event, wherein the first signal subsequently drops below the first threshold value; and

generating an output signal, wherein the magnitude of the output signal is based on an event time equal to the time elapsed between the first event and the second event during a fixed time after the second event.

19. The method of claim 18, wherein the magnitude of the output signal is zero other than during the fixed time after the second event.

20. The method of claim 18, wherein the magnitude of the output signal during the fixed time after the second event is based on the difference of the event time and a predetermined value.