Associative neuron in an artificial neural network

• US 6,625,588 B1
• Filed: 09/24/1999
• Issued: 09/23/2003
• Est. Priority Date: 03/26/1997
• Status: Expired due to Term

• Current Assignee: Nokia Solutions and Networks US LLC (Nokia Corporation)
• Original Assignee: Nokia Oyj (Nokia Corporation)
• Inventors: Haikonen, Pentti
• Primary Examiner(s): Patel, Ramesh
• Assistant Examiner(s): Hirl, Joseph P.
• Application Number:
  US09/381,825
• Time in Patent Office: 1,460 Days
• Field of Search: 706/33, 706/34, 706/37, 706/38, 706/41, 706/18
• US Class Current: 706/33
• CPC Class Codes: G06N 3/063 using electronic means

Abstract

An associative artificial neuron and method of forming output signals of an associative artificial neuron includes receiving a number of auxiliary input signals; forming from the auxiliary input signals a sum weighted by coefficients and applying a non-linear function to the weighted sum to generate a non-linear signal. The neuron and method further include receiving a main input signal and forming, based on the main signal and the non-linear signal, the function S OR V, which is used to generate a main output signal, and at lest one of three logical functions S AND V, NOT S AND V, and S AND NOT V. The at least one logical function is used to generate an additional output signal for the associative artificial neuron.

Claims

1. A method of forming output signals of an associative artificial neural network, the method comprising;

receiving a number of auxiliary signals;

forming a corresponding weighting coefficient for each auxiliary signal;

forming from the auxiliary signals a sum weighted by the corresponding coefficients;

applying a non-linear function to the weighted sum to generate a non-linear signal V;

receiving a main signal S, which can be associated with the auxiliary signals such that the weighting coefficient of each auxiliary signal is increased when the main signal S and said auxiliary signal are simultaneously active; and

forming, based on the main signal S and the non-linear signal V, a function S OR V, which is used to generate a main output signal, and at least one of three logical functions, including S AND V, NOT S AND V, and S AND NOT V, and using said at least one logical function to generate an additional output signal for the associative artificial neural network.

2. The method of claim 1, wherein the non-linear function is a threshold function, and the non-linear signal V is obtained by the threshold function has a first state and a second state.

3. The method of claim 1, wherein the non-linear function is a step function with more than two steps.

4. The method of claim 1, wherein the non-linear function is a continuous function.

5. The method of claim 1, wherein the main output signal has a first state and a second state and the method further comprises setting an upper limit to a length of time that the main output signal is in the second state.

6. The method of claim 5, further comprising setting a lower limit to a length of time that the main output signal remains in the first state after having been in the second state.

7. The method of claim 1, further comprising adjusting a learning of the associative artificial neuron in response to an external control signal.

8. The method of claim 7, wherein adjusting the learning of the associative artificial neuron has two states, whereby the learning of the associative artificial neuron is either allowed or inhibited.

9. The method of claim 8, further comprising forcing at least one output signal to a predetermined state in response to an external inhibiting signal.

10. The method of claim 7, further comprising forcing at least one output signal to a predetermined state in response to an external inhibiting signal.

11. The method of claim 1, further comprising forcing at least one output signal to a predetermined state in response to an external inhibiting signal.

12. An associative artificial neuron comprising:

means for receiving a number of auxiliary signals and forming a corresponding coefficient for each auxiliary signal;

means for forming from the auxiliary signals a sum weighted by the corresponding coefficients;

means for applying a non-linear function to the weighted sum to generate a non-linear signal V;

means for receiving a main signal S, which can be associated with the auxiliary signals such that said weighting coefficient for each auxiliary signal is increased when the main signal S and said auxiliary signal are simultaneously active; and

means for forming, based on the main signal S and the non-linear signal V, a function S OR V, which is used to generate a main output signal, and at least one of three logical functions, including S AND V, NOT S AND V, and S AND NOT V, and using said at least one logical function to generate an additional output signal for the associative artificial neuron.

13. An associative artificial neuron according to claim 12, wherein said non-linear function is a threshold function and the non-linear signal V is obtained by the threshold function has a first state and a second state.

14. An associative artificial neuron according to claim 12, wherein the non-linear function is a step function with more than two steps.

15. An associative artificial neuron according to claim 12, wherein the non-linear function is a continuous function.

16. An associative artificial neuron according to claim 12, wherein said main output signal has a first state and a second state and the associative artificial neuron further comprises means for setting an upper limit to a length of time that said main output signal is in the second state.

17. An associative artificial neuron according to claim 16, further comprising means for setting a lower limit to a length of time that said main output signal remains in the first state after having been in the second state.

18. An associative artificial neuron according to claim 12, further comprising means for adjusting a learning of the associative artificial neuron in response to an external control signal.

19. An associative artificial neuron according to claim 18, wherein the means for adjusting the learning of the associative artificial neuron has two states, whereby the learning of the associative artificial neuron is either allowed or inhibited.

20. An associative artificial neuron according to claim 18, further comprising means for forcing at least one output signal to a predetermined state in response to an external inhibiting signal.

21. An associative artificial neuron according to claim 12, further comprising means for forcing at least one output signal to a predetermined state in response to an external inhibiting signal.

22. An associative artificial neuron according to claim 19, further comprising means for forcing at least one output signal to a predetermined state in response to an external inhibiting signal.