Multilayer modular computer system

FIELD: information technology.

SUBSTANCE: multilayer modular computer system has several layers, including a neural network layer, a transport layer and a processor layer, wherein the transport layer contains network controller-router modules, the processor layer contains processor modules, and all the said modules have multiple inputs and outputs connected to each other and connected to the inputs and outputs of the system. The processor modules train neural network domain modules.

EFFECT: high decision speed, possibility of grafting layers and modules in each layer during operation of the system with a complex task, high reliability of the computer system.

3 cl, 1 dwg

 

The claimed device is multi-layered modular computer system refers to computing and can be used to create complex automated control systems and systems with artificial intelligence, including machine control, process engineering, systems of automatic control of aircraft and space technology, etc.

Known examples of the various modular computing systems.

For example, the Cray T3D computer, which is a supercomputer with massive parallelism and distributed memory [1]. It consists of two types of components: compute nodes and the switching network. The switching network of the Cray T3D computer produces a three-dimensional lattice connecting the network routers nodes in the three spatial directions. The link between two adjacent nodes is implemented using two unidirectional data channels that allow simultaneous communication in opposite directions. A disadvantage of this device is low, the actual performance when solving problems that require a large number megauploadcom shipments.

Known multiprocessor system consisting of modules [2]. Module multi-processor systems used to build multiprocessor systems, who will win the group macroprocessors, performing large mathematical operations, the group Multicontroller distributed memory, providing high-speed exchange of information between RAM and microprocessore and parallel-pipelined data processing, matrix switch that provides direct spatial connection between all system components, and information input devices are connected with bidirectional inputs/outputs memory and bidirectional inputs/outputs block Multicontroller distributed memory.

Known multi-layer neural network [3], which contain several layers of neurons. Neurons have many inputs and many outputs. The inputs and outputs of the neurons of each layer are connected respectively to the inputs and outputs of the neural network and the outputs and inputs of the neurons of other layers. The inputs and outputs of the neurons of one layer are not connected.

Known utility model "Neural network" and "Domain neural network [4, 5]. In neural network [4] there are neurons and switches. Neurons have one input and one output, and perform the function of information processing and decision making. The inputs and outputs of the neuron connected to the inputs and outputs of the switches. Switch the neural network consists of a table of relationships of neurons constituting the network fragment, and a device that transmits information between the at neurons on the basis of this table. The inputs and outputs of the switch are connected to the inputs and outputs of other switches and inputs and outputs of the neurons in the network.

Fragments of the network with interconnected neurons and switches combined into domains [5]. The domains are connected to the switches at a higher level. The domain is the gateway. The domain has one input and one output, which are formed respectively by the external input and external output gateway. The gateway has one external input and one external output, as well as many internal inputs and internal outputs. The gateway contains a table locking that specifies the connection of an external input and external output gateway respectively with the internal inputs and outputs of the gateway device, receiving and transmitting information on the basis of this table. Internal inputs and internal outputs of the gateway connected respectively to the inputs and outputs of the internal switches, neurons and domains lower level included in the network fragment, which forms a domain.

Closest to the technical nature of the claimed device is a useful model "Modular computing system [6].

The device is modular computing system /HP-6/, where each module contains several connectors, network controller, router, gateway, and functional blocks, and the connector is used for mutual connection module is, the gateway module has one external input and one external output, as well as many internal inputs and outputs connected respectively to the outputs and inputs mentioned functional blocks, characterized in that the inputs and outputs of the controller router connected to the connectors, and external input and external output of the gateway.

The device is modular computing system [6], characterized in that all modules have the same shape and size, and the module is made as three-dimensional geometric figure in which all sides have the shape of an equilateral n - gon, on the surface of each face have connectors for connecting the modules together, and also for reception and transmission of information and the supply of the supply voltage.

The claimed device is multi-layered modular computer system allows to implement complex automatic control systems and artificial intelligence on the basis of the large domain of neural networks.

The technical result of the claimed device is:

- creating complex modular computing systems based on large domain of neural networks;

- increased speed of decision-making;

- reduction of production costs due to the unification of nomenclature modules;

- the ability to build layers and modules in each layer during operation of the system PR is the complication of the task.

- improving the reliability of the computing system by introducing modules and layers.

Claim 1. The device multi-layered modular computer system, where there are several layers, including the layer of the neural network, the transport layer and the processing layer, and transport layer contains modules network controllers, routers, CPU layer contains processor modules, and these modules have many inputs and many outputs, interconnected and connected to the inputs and outputs of the system,

where the processor modules perform the functions of the interface with the external environment, the initialization of the computing system, routing information transfer and reconfiguration, memory function and store information about the state of the system and its modules, the disaster recovery system in case of failure,

and modules network controllers, routers perform the functions of receiving and transmitting information between the layers and modules of the system,

characterized in that the layer of the neural network consists of modules domains, where the unit domain of the neural network has one external input and one external output, which is connected to the inputs and outputs of one module network controller router.

2. The device according to claim..1, characterized in that the processor modules perform the functions of training modules dominiando network.

P.3. The device according to claim 2, characterized in that each layer of the system, in turn, is built from the individual sub-layers according to the same principles as the device of the computing system as a whole.

The scheme of the claimed device multi-layered modular computer system shown in the drawing.

The device multi-layered modular computer system (the drawing), which consists of interconnected layers n1, n2, n3.

The number of layers of the modular computing system can be arbitrarily large (but not less than three) depending on the complexity of the task and the reliability requirements. In the process of operation to a modular computing system may include additional layers as needed.

Each layer contains unified computing modules of the same type (4, 5, 6). The inputs and outputs of the modules are interconnected and also connected to the inputs and outputs of the system. For mutual connection of the modules to each other and to transmit and receive data between the module connectors are used.

In multi-layered modular computer system has several layers, including the layer of the domains of the neural network n1, the transport layer n2 and processor layer n3. Each layer contains unified modules of the same type, which have inputs and outputs. Layer n1 contains modules domains of the neural network threeparty layer n2 contains the modules of the network controllers, routers 5. The processor layer n3 contains processor modules 6. The inputs and outputs of the modules are connected to the inputs and outputs of the modules of the same and of the other layers, and inputs and outputs of the system. If necessary, backup modules, and when the complication and expansion of the functionality of the device adds new modules or new layers.

The device operates as follows. Module domain of the neural network 4 is composed of neurons and switches, and is used as a storage device and information processing. Module domain 4 has one input and one output, which is connected to the input and output of one module network controller router 5. Module network controller router 5 has many inputs and outputs that are connected to the inputs and outputs of the modules of the 4 domains, other controllers, routers 5 and processor modules 6. The inputs and outputs of the processor modules 6 are connected to the inputs and outputs of the modules of the network controllers, routers 5 and other processor elements 6, as well as to the inputs-outputs of the computing system (i/o 1 ... i/o n3). Module network controller router 5 performs the function of transmitting information between modules, and the route of transmission processor sets the layer.

Information is passed between modules in the form of data packets, which includes the address of the layer-send the I and module-sender, address layer-recipient module of the receiver. Module network controller router 5 performs reception and transmission of information in accordance with the addresses of the layers and modules of the system.

Network controller-router determines the address layer and the module for which you received the information. If this information is passed to the layer, the controller router will forward it to the module with the address of the recipient or another controller router. If this information is transmitted to another layer, the controller router passes it to the transport layer, closest to the module to the recipient. For routing information between modules uses an algorithm similar to message routing in a computer network (e.g. the Internet).

The processor module may be implemented as a digital or analog device. If the processor module is implemented as an analog device, all the outputs must have an analog-to-digital Converter (ADC), and inputs - analog Converter (DAC).

Processor modules perform the functions of the interface with the external environment, the initialization of the computing system, the routing information in the system, neural network training and recovery failure of the entire system, the individual layers or modules.

Processor modules perform the functions about which the respective domains of the neural network. When training module domain neural network memory CPU module the table is formed of switching inputs and outputs of the neurons within a domain, and table switching inputs and outputs of modules domains within a layer.

Table switching inputs and outputs of the neurons within the domain contains links and the weights of the outputs of the neurons included in this domain.

In table switching inputs and outputs domains inside layer recorded similar information, but for a higher level of mutual communication and weights module outputs domains.

After learning domains of the neural network of the referenced table is copied from the memory modules processor layer switches domains of the neural network. Based on this information, the processor layer defines the transmission path information, and then transmits the route to the transport layer, which implements it. The transport layer provides information transfer between domains of the neural network.

Improving fault tolerance of multilayer modular system is possible due to a number of backing layers or modules in each layer. Failure of any module domain neural network there is no need for re-training. Upon cancellation of a domain or a network controller, router table switching domain is copied from the memory% is scornage element in the switch backup domain and then being re-routed. This ensures high reliability of the device.

The processing layer performs control functions transport layer (including routing and reconfiguration), as well as the functions of the interface with the external environment and memory, which contains full information about the modules of the computing system.

The modular system can consist of a large number of layers, each layer of the system, in turn, can be built from sub-layers according to the same principles as the device of the computing system as a whole.

Sources of information

1. Computing processors and systems. Ed. Genmarshal. Issue 7. - M.: Nauka. The main edition of physico-mathematical literature, 1990.

2. Levin I.I., Vinevskaja LI Module multiprocessor system. Application No. 2004136937/09, the filing date 16.12.2004, IPC G06F 15/16.

3. Galushkin A.I. Theory of neural networks. KN. 1: - M: IPGR, 2000. - 416 S.: ill (Neurocomputers and their application).

4. The tavern I.S., Sukhanova, NV Neural network. Patent for useful model №66831, publ. 27.09.2007, bull. No. 27.

5. The tavern I.S., Sukhanova, NV Domain neural network. Patent for useful model №72084, publ. 27.03.2008, bull. No. 9.

6. The tavern I.S., Sukhanova, NV Modular computing system. Patent for useful model №75247, publ. 27.07.2008.

1. The device is a multilayer modular computing the system, where there are several layers, including the layer of the neural network, the transport layer and the processing layer, and transport layer contains modules network controllers, routers, CPU layer contains processor modules, and these modules have many inputs and many outputs, interconnected and connected to the inputs and outputs of the system, where the processor modules perform the functions of the interface with the external environment, the initialization of the computing system, routing information transfer and reconfiguration, memory function and store information about the state of the system and its modules, the disaster recovery system in case of failure, and the modules of the network controllers, routers perform the functions of receiving and transmitting information between the layers and modules of the system, characterized in that the layer of the neural network consists of modules domains, where the unit domain of the neural network has one external input and one external output, which is connected to the inputs and outputs of one module network controller router.

2. The device of claim 1, wherein the processor modules perform the functions of training modules domains of the neural network.

3. The device according to claim 2, characterized in that each layer of the system, in turn, is built from the individual sub-layers according to the same principles as the device will calculate the school system as a whole.



 

Same patents:

FIELD: information technologies.

SUBSTANCE: invention may be used for building of modular neural computers, which function in symmetrical system of residual classes. Stated neuron network comprises unit of neuron network of end ring of senior coefficient generation for generalised positional system of numeration, unit of polarity shift, unit of error detection, buses "with errors" and "without errors".

EFFECT: reduced hardware complexity.

3 dwg

FIELD: physics; computer engineering.

SUBSTANCE: present invention pertains to neurocomputers. The device has a unit for storing a binary input signal, a logic AND-OR circuit, internal memory unit, unit for generating the output string of codes, a generator of synchronising pulses, control unit, a unit for selecting duration and extracting information, analysis block and a corrector unit.

EFFECT: increased rate of operation, providing for the possibility of distinguishing change in state of processed signals, increased noise immunity, possibility of making super-complex neural networks, and simplification of design.

9 cl, 1 dwg

FIELD: modular neuro-computing systems.

SUBSTANCE: neuron network contains input layer of neurons, at inputs of which residuals of number being divided are received through system of modules, (n-1) neuron networks of finite ring for addition, (n-1) neuron networks of finite ring for multiplication, neuron network for expanding a tuple of numerical system of residues, and as output of neuron network for dividing numbers represented in system of residual classes are outputs of neuron network of finite ring for multiplication and output of neuron network for expansion of tuple of numerical system of residues.

EFFECT: expanded functional capabilities, increased speed of division, reduced volume of equipment.

1 dwg

FIELD: neuron-like computing structures, possible use as processor for high speed computer systems.

SUBSTANCE: device contains artificial neuron network composed of analog neurons, at least one controllable voltage block, a group of long neuron-like nonlinear communication units, each one of which contains serially connected circuit for synchronization and selection of radio impulse envelope, auto-generator with self-suppression circuit, a length of coaxial line, realizing functions of antenna, additional circuit for synchronization and selection of radio-impulse envelope.

EFFECT: increased information processing speed due to increased paralleling degree of computing processes.

2 dwg

FIELD: neuro-cybernetics, possible use in artificial neuron networks for solving various problems of logical processing of binary data.

SUBSTANCE: method for realization of logical nonequivalence function by neuron with two inputs is based on multiplication of input signals with corresponding weight coefficients and summing them, after that the total is transformed in activation block firstly by quadratic transfer function, and then by threshold function at neuron output.

EFFECT: realization by one neuron of first order of logical nonequivalence function of two variables.

5 dwg, 1 tbl

FIELD: computer engineering, possible use in modular neuro-computer systems.

SUBSTANCE: in accordance to invention, neuron network contains input layer, neuron nets of finite ring for determining errors syndrome, memory block for storing constants, neuron nets for computing correct result and OR element for determining whether an error is present.

EFFECT: increased error correction speed, decreased amount of equipment, expanded functional capabilities.

1 dwg, 3 tbl

Neuron-like element // 2295769

FIELD: cybernetics, possible use as a cell for neuron networks.

SUBSTANCE: neuron-like element may be used for realization on its basis of neuron network for solving problems of estimation of functioning of complicated open systems, estimation of degree of optimality of various solutions by ensuring possible construction of model of researched system, both hierarchical and recurrent, with consideration of varying original and working condition of its elements and variants of their functioning, during modeling taking into consideration the level of self-sufficiency of neuron-like elements, susceptibility to effect of external signals, type and errors of setting of their parameters and parameters of input signals, and also provision of given precision of self-teaching of neuron network. Device contains input block, block for setting and normalizing weight coefficients, block for computing parameters of input signals, adder, signals share limiter, block for computing input part of condition, block for setting internal state, block for computing internal part of distance, block for counting distance, memory block, analyzer of state change value, block for determining precision of self-teaching of neuron network, block of determined dependencies, switch, output block, control block, random numbers generator.

EFFECT: creation of neuron-like element.

2 cl, 1 dwg

The invention relates to the field of computer engineering and can be used in neural networks

The invention relates to the field of computer engineering and can be used in neural computers

Neuron-like element // 2295769

FIELD: cybernetics, possible use as a cell for neuron networks.

SUBSTANCE: neuron-like element may be used for realization on its basis of neuron network for solving problems of estimation of functioning of complicated open systems, estimation of degree of optimality of various solutions by ensuring possible construction of model of researched system, both hierarchical and recurrent, with consideration of varying original and working condition of its elements and variants of their functioning, during modeling taking into consideration the level of self-sufficiency of neuron-like elements, susceptibility to effect of external signals, type and errors of setting of their parameters and parameters of input signals, and also provision of given precision of self-teaching of neuron network. Device contains input block, block for setting and normalizing weight coefficients, block for computing parameters of input signals, adder, signals share limiter, block for computing input part of condition, block for setting internal state, block for computing internal part of distance, block for counting distance, memory block, analyzer of state change value, block for determining precision of self-teaching of neuron network, block of determined dependencies, switch, output block, control block, random numbers generator.

EFFECT: creation of neuron-like element.

2 cl, 1 dwg

FIELD: computer engineering, possible use in modular neuro-computer systems.

SUBSTANCE: in accordance to invention, neuron network contains input layer, neuron nets of finite ring for determining errors syndrome, memory block for storing constants, neuron nets for computing correct result and OR element for determining whether an error is present.

EFFECT: increased error correction speed, decreased amount of equipment, expanded functional capabilities.

1 dwg, 3 tbl

FIELD: neuro-cybernetics, possible use in artificial neuron networks for solving various problems of logical processing of binary data.

SUBSTANCE: method for realization of logical nonequivalence function by neuron with two inputs is based on multiplication of input signals with corresponding weight coefficients and summing them, after that the total is transformed in activation block firstly by quadratic transfer function, and then by threshold function at neuron output.

EFFECT: realization by one neuron of first order of logical nonequivalence function of two variables.

5 dwg, 1 tbl

FIELD: neuron-like computing structures, possible use as processor for high speed computer systems.

SUBSTANCE: device contains artificial neuron network composed of analog neurons, at least one controllable voltage block, a group of long neuron-like nonlinear communication units, each one of which contains serially connected circuit for synchronization and selection of radio impulse envelope, auto-generator with self-suppression circuit, a length of coaxial line, realizing functions of antenna, additional circuit for synchronization and selection of radio-impulse envelope.

EFFECT: increased information processing speed due to increased paralleling degree of computing processes.

2 dwg

FIELD: modular neuro-computing systems.

SUBSTANCE: neuron network contains input layer of neurons, at inputs of which residuals of number being divided are received through system of modules, (n-1) neuron networks of finite ring for addition, (n-1) neuron networks of finite ring for multiplication, neuron network for expanding a tuple of numerical system of residues, and as output of neuron network for dividing numbers represented in system of residual classes are outputs of neuron network of finite ring for multiplication and output of neuron network for expansion of tuple of numerical system of residues.

EFFECT: expanded functional capabilities, increased speed of division, reduced volume of equipment.

1 dwg

FIELD: physics; computer engineering.

SUBSTANCE: present invention pertains to neurocomputers. The device has a unit for storing a binary input signal, a logic AND-OR circuit, internal memory unit, unit for generating the output string of codes, a generator of synchronising pulses, control unit, a unit for selecting duration and extracting information, analysis block and a corrector unit.

EFFECT: increased rate of operation, providing for the possibility of distinguishing change in state of processed signals, increased noise immunity, possibility of making super-complex neural networks, and simplification of design.

9 cl, 1 dwg

FIELD: information technologies.

SUBSTANCE: invention may be used for building of modular neural computers, which function in symmetrical system of residual classes. Stated neuron network comprises unit of neuron network of end ring of senior coefficient generation for generalised positional system of numeration, unit of polarity shift, unit of error detection, buses "with errors" and "without errors".

EFFECT: reduced hardware complexity.

3 dwg

FIELD: information technology.

SUBSTANCE: multilayer modular computer system has several layers, including a neural network layer, a transport layer and a processor layer, wherein the transport layer contains network controller-router modules, the processor layer contains processor modules, and all the said modules have multiple inputs and outputs connected to each other and connected to the inputs and outputs of the system. The processor modules train neural network domain modules.

EFFECT: high decision speed, possibility of grafting layers and modules in each layer during operation of the system with a complex task, high reliability of the computer system.

3 cl, 1 dwg

FIELD: physics.

SUBSTANCE: neuron simulation method is based on calculation of squares of Euclidean distance from the input vector to each of 2n vertices of a unit n-dimensional cube in weighting units, and multiplication of values inverse to these distance values with components of the target vector respectively, and then summation in an adder and conversion in the activation unit through an activation function.

EFFECT: possibility of simulating a neuron of any given Boolean function from a complete set of from n variables.

6 dwg, 1 tbl

FIELD: information technology.

SUBSTANCE: in an ophthalmic-microsurgical computer local area network for vitreoretinal operations, formatting devices are in form of a radial-annular structure consisting of a single set of automated workstations (AWS), which synchronously or asynchronously functioning, processing, converting, transmitting, analysing, synthesising hierarchical structures of an artificial neural network: diagnosis AWS (DAWS), ophthalmic-microsurgical AWS (OMAWS), subsequent operation stages AWS (SOSAWS), component AWS (CAWS), surgeon's operating unit (SOUAWS), with opposite forward and reverse flow of information in between, where each AWS has at least one neural circuit, interconnected identification units (IU), an interpolation unit (INU), an extrapolation unit (EU), which are the neural network converting and transmitting elements (NNCTE), a decision unit (DU), which is the neural network analysis and synthesis element (NNASE).

EFFECT: simultaneous improvement of accuracy of determination and quality of identifying diagnoses, determining indications for conducting operations, high selectivity when conducting operations, accuracy in determining the sequence of operations, simulating operations, accuracy in choosing the anaesthetic method, accuracy of providing implants and expendable materials, optimisation of flow of information and necessities during vitreoretinal ophthalmic-microsurgical operations.

1 dwg

Up!