US 8,027,942 B2 · Method and circuits for associating a complex operator to each component of an input pattern presented to an artificial neural network

Abstract

The method and circuits of the present invention aim to associate a complex component operator (CC_op) to each component of an input pattern presented to an input space mapping algorithm based artificial neural network (ANN) during the distance evaluation process. A complex operator consists in the description of a function and a set of parameters attached thereto. The function is a mathematical entity (either a logic operator e.g. match(Ai,Bi), abs(Ai−Bi), . . . or an arithmetic operator, e.g. >, <, . . . ) or a set of software instructions possibly with a condition. In a first embodiment, the ANN is provided with a global memory, common for all the neurons of the ANN, that stores all the CC_ops. In another embodiment, the set of CC_ops is stored in the prototype memory of the neuron, so that the global memory is no longer physically necessary. According to the present invention, a component of a stored prototype may now designate objects of different nature. In addition, either implementation significantly reduces the number of components that are required in the neurons, therefore saving room when the ANN is integrated in a silicon chip.

Claims7

Dependent claims

An artificial neural network having a plurality of neurons adapted to associate a complex component operator to each of a plurality of components of an input pattern in the process of evaluating a final distance between the input pattern and a plurality of prototypes stored in the neurons of the artificial neural network comprising:

data-in transmitting means for applying input pattern components to each of the neurons;
storage means to store and apply a plurality of complex component operators to each of the neurons via a bus, wherein each of the complex component operators has a function and at least one parameter attached thereto;
said plurality of neurons in which each of the neurons has a prototype memory to store a plurality of components of a prototype and distance evaluation means connected to said prototype memory, to said storage means via said bus and to said data-in transmitting means;
addressing means for applying the input pattern components, the corresponding one of the plurality of complex component operators associated therewith and the corresponding stored prototype component to said distance evaluation means in each of the neurons;
means for applying a distance norm to the distance evaluation means;
wherein said distance evaluation means are adapted to perform an elementary distance calculation between the corresponding components of the input pattern and the stored prototype using the corresponding complex component operator associated thereto for each component, said complex component operator being obtained by decoding of said function and processing of said at least one parameter and said distance evaluation means is also adapted to combine said calculated elementary distance to determine the final distance using said distance norm.

3 more elements

Dependencies are inferred from the claim text.

Method and circuits for associating a complex operator to each component of an input pattern presented to an artificial neural network

Life of this patent

Method and circuits for associating a complex operator to each component of an input pattern presented to an artificial neural network

Life of this patent

Abstract

Claims7