Nokia Corporation

Intellectual Ventures I LLC

Nokia Oyj

Pentti Haikonen

The invention relates to circuit elements and computing networks for resolving logical entanglement, in which the allowed logical value of a variable in a set of variables depends on the logical values of the other variables in the set. A circuit element according to the invention comprises two or more logically entangled bi-directional terminals, wherein each bi-directional terminal can assume any one of three logical states, which are a logical true state, a logical false state, and an indefinite state, in which state the bi-directional terminal accepts one of the logical true and logical false states as an external input from an external source. An entanglement logic resolves the logical state of the bi-directional terminals according to a predetermined set of logical entanglement rules between the bi-directional terminals.

Circuit elements and parallel computational networks with logically entangled terminals

The invention relates to a method for determining a weight coefficient of an artificial associative neuron synapse, where the synaptic weight coefficient is determined on the basis of the temporal average of a product of two signals. The method comprises the steps of taking temporal samples from the product of said two signals at such moments when one of the signals starts to deviate from zero, feeding said samples into such a shift register chain, from where a predetermined number of said samples taken at previous moments are continuously available, deducing on the basis of said samples taken at previous moments, whether a value deviating from zero is to be set as the synaptic weight coefficient. The invention also relates to an artificial associative neuron synapse.

Artificial associative neuron synapse

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.

Growing community of inventors

Helsinki, Finland

Pentti Haikonen