The paper describes an approach for implementing Boolean neuralnetworks on silicon. The hardware is based on a custom designed Field Programmable Logic Device (FPLD) which integrates 'synapses' and 'neuro...
详细信息
The paper describes an approach for implementing Boolean neuralnetworks on silicon. The hardware is based on a custom designed Field Programmable Logic Device (FPLD) which integrates 'synapses' and 'neurons' and allows random access to the weights during training. networks are realised from arrays of the neural chip which are assembled on ceramic as Multi-Chip Modules (MCM) to provide expandability and flexibility. The hardware provides parallel computation of the 'neuron' outputs and promises significantly improved performance compared to purely software approaches.
A VLSI Implementation of an Artificial neural Network using a single n-channel MOS transistor per synapse is investigated. The simplicity of the design is achieved by using pulse width modulation to represent neural a...
详细信息
A VLSI Implementation of an Artificial neural Network using a single n-channel MOS transistor per synapse is investigated. The simplicity of the design is achieved by using pulse width modulation to represent neural activity and a novel technique for manipulating synaptic weights. A multi layer perceptron network built in hardware gives good results for a simple classification task.
The reconfigurability of certain field programmable gate arrays(FPGAs) has shown their advantage of flexibility in digital system design. With the availability of greater density and high speed of FPGAs, the ability t...
详细信息
The reconfigurability of certain field programmable gate arrays(FPGAs) has shown their advantage of flexibility in digital system design. With the availability of greater density and high speed of FPGAs, the ability to realise special purpose processors will become possible. In this paper, we present our research work of implementing specific trained neural network in Xilinx XC4000 series FPGAs for portable digital system prototype in heart disease classification process.
The competence of the HyperNet system was evaluated. In addition, an advanced path able to satisfy the standards of more complicated, extensive, real-time uses were discussed. The HyperNet system integrates probabilis...
详细信息
The competence of the HyperNet system was evaluated. In addition, an advanced path able to satisfy the standards of more complicated, extensive, real-time uses were discussed. The HyperNet system integrates probabilistic, RAM-based, feed forward design and employs a conventional VLSI IC with on-board Reward-Penalty learning program. The system also provides opportunity for inexpensive Ultra Large Scale and Wafer Scale Integration.
Artificial neuralnetworks have shown good capabilities in medical diagnostic applications. They offer the advantage that they are able to learn the representation by examples, which is of great benefit when the natur...
详细信息
Artificial neuralnetworks have shown good capabilities in medical diagnostic applications. They offer the advantage that they are able to learn the representation by examples, which is of great benefit when the nature of the process is unknown or is difficult to characterise. On the other hand, the hardware implementation of the parallel network structure can dramatically improve the network efficiency. In this paper, A hardware implementation of neural network based ballistocardiogram (BCG) classification system with field programmable gate arrays(FPGAs) technology is presented. The specific trained neural network is implemented in Xilinx XC4000 series FPGAs.
The paper proposes a novel adaptive three-phase autoreclosure technique for double circuit systems using a neural network approach. Based on the investigation of digital simulation of various types of fault on such sy...
详细信息
The paper proposes a novel adaptive three-phase autoreclosure technique for double circuit systems using a neural network approach. Based on the investigation of digital simulation of various types of fault on such systems, some salient features are summarized and extracted which are then used as the inputs of neuralnetworks. A three-layer neural network is constructed, trained and tested. The results indicate that the proposed approach is very reliable.
The paper deals with new developments of interpolating memories as the basic element of learning control and with their possible application. It discusses learning control, interpolating memories, characteristic manif...
详细信息
The paper deals with new developments of interpolating memories as the basic element of learning control and with their possible application. It discusses learning control, interpolating memories, characteristic manifolds for automotive control, and possible future developments.< >
Hypersonic aircraft require a high degree of system integration. Design tools are needed that can provide rapid, accurate calculations of complex fluid flow. Existing methods are slow. The goal of this project was to ...
详细信息
Hypersonic aircraft require a high degree of system integration. Design tools are needed that can provide rapid, accurate calculations of complex fluid flow. Existing methods are slow. The goal of this project was to apply neuralnetworks to the calculation of fluid flow and heat transfer in a heat exchanger panel for the National AeroSpace Plane (NASP).< >
In this contribution, neural concepts and methods for control an their use in industrial applications are discussed and illustrated. Even though there exist numerous conventional approaches for solving control tasks, ...
详细信息
In this contribution, neural concepts and methods for control an their use in industrial applications are discussed and illustrated. Even though there exist numerous conventional approaches for solving control tasks, their realization in practice frequently proves to be very difficult. The author pursues several approaches to using neuralnetworks in the context of nonlinear control tasks. In identification, networks are trained to model the dynamics of an unknown nonlinear plant. The model provides the basis for controller design or system diagnosis. In robot control, networks are trained to model the inverse dynamics of the robot. The inverse model is used to linearize the system which is then accessible for the well-established tools of linear control theory. In another context, a network is used as a nonlinear trainable controller.< >
The training speed of batch backpropagation using steepest descent, conjugate gradient and quasi-Newton algorithm for a feedforward neural network are compared. Results illustrating the advantages of the Hessian based...
详细信息
The training speed of batch backpropagation using steepest descent, conjugate gradient and quasi-Newton algorithm for a feedforward neural network are compared. Results illustrating the advantages of the Hessian based techniques are given and issues affecting speed discussed.< >
暂无评论