THE PYGMALION NEURAL NETWORK PROGRAMMING ENVIRONMENT

We here describe SNNS, a neural network simulator for Unix workstations that we have been developing at the Universitat Stuttgart, Germany. Our network simulation environment is a tool to generate, train, test and visualize artificial neural networks. The simulator consists of three major components: a simulator kernel that operates on the internal representation ofthe neural networks, a graphical user interface based on X-Windows to interactively construct and change small neural nets,:ind a compiler to generate the internal representation of large neural networks from a high level network description language.Keywords: connectionism, neural networks, network simulators, network description language 1. INTRODUCTION AND MOTIVATION SNNS (Stuttgart Neural Network Simulator) is an efficient and portable neural network simulation environment for Unixworkstations that we have been developing at the Institute for Parallel and Distributed High Performance Systems, Universitht Stuttgart, Germany. It is an efficient software tool to generate, train, test and visualize artificial neural networks. The

To understand the mechanism of information processing by a biological neural network, computer simulation of a large-scale spiking neural network is an important method. However, because of a high computation cost of the simulation of a large-scale spiking neural network, the simulation requires high performance computing implemented by a supercomputer or a computer cluster. Recently, hardware for parallel computing such as a multi-core CPU and a graphics card with a graphics processing unit (GPU) is built in a gaming computer and a workstation. Thus, parallel computing using this hardware is becoming widespread, allowing us to obtain powerful computing power for simulation of a large-scale spiking neural network. However, it is not clear how much increased performance the parallel computing method using a new GPU yields in the simulation of a large-scale spiking neural network. In this study, we compared computation time between the computing methods with CPUs and GPUs in a simulation of neuronal models. We developed computer programs of neuronal simulations for the computing systems that consist of a gaming graphics card with new architecture (the NVIDIA GTX 1080) and an accelerator board using a GPU (the NVIDIA Tesla K20C). Our results show that the computing systems can perform a simulation of a large number of neurons faster than CPU-based systems. Furthermore, we investigated the accuracy of a simulation using single precision floating point. We show that the simulation results of single precision were accurate enough compared with those of double precision, but chaotic neuronal response calculated by a GPU using single precision is prominently different from that calculated by a CPU using double precision. Furthermore, the difference in chaotic dynamics appeared even if we used double precision of a GPU. In conclusion, the GPU-based computing system exhibits a higher computing performance than the CPU-based system, even if the GPU system includes data transfer from a graphics card to host memory.

The application of neural network in cryptographic known plaintext attack is a novel cryptanalysis method. Only with known ciphertext, the restoration of plaintext is a valuable part of cryptanalysis. By putting plaintext and ciphertext into model, the network simulates encryption algorithm without the key and output the corresponding plaintext. In this paper, we focus on the plaintext restoration of Blowfish based on neural network. We design a plaintext restoration scheme and perform the experiment in open source database. The result shows that the neural network simulator could fit the dataset with an error rate lower than 0.04 when the size of dataset exceeds 219•.

Neural networks provide the opportunity to estimate specific components of engineering problems. They are decomposed complex problems into different parts. Thus, it can be easy to compete with each of them through neural networks. In this paper, it was purposed to estimate the average speed of a 6-line road’s cross-section by observed traffic variables, such as numbers of vehicles and occupancy values, using radial basis function neural network (RBFNN), generalized regression neural network (GRNN) and the feed-forward back propagation neural network (FFBPNN) models. A comparison was fulfilled between different neural networks and checked against multivariate linear regression (MVLR), a conventional statistical model. After each simulation of neural networks, results show that different forecasts were obtained under the same conditions. The best forecasting is made by FFBPNN, GRNN, and RBFNN, respectively. When compared with multivariate linear regression (MVLR), FFBPNN performs better than MVLR, but GRNN and RBFNN perform lower than it.

The Sixth International Symposium on Neural Networks (ISNN 2009)

This paper describes the design, implementation and testing of the software tool designated as UV-SRNA-PDA (Universidad del Valle’s artificial neural network simulator for a personal digital assistant), which is oriented to artificial neural networks simulation and complex industrial processes identification. This application works on a PDA (personal digital assistant) Palm T5 using a customized data acquisition system. Two kinds of artificial neural networks were implemented: perceptron and multi-layer perceptron (MLP) using as learning algorithms the following: backpropagation, descendent gradient, variable learning rate and momentum. To test the proposed system, first and second order systems were selected (only the latter is reported here), finding their neural models and validating their results with MATLAB and UV-SRNA 2.0 (PC version of UV-SRNA-PDA). These tests achieved an average training error of 5.62 × 10-3 ± 3.55 × 10-4 and an average validating error of 4.56 × 10-3 ± 5.95 × 10-4. In both cases, the results were better or comparable with those from other software simulation tools. However, the typical training time on the UV-SRNA-PDA was 900 s compared to 3 s in MATLAB and 8 s in UV-SRNA 2.0.

The wide applicability of neural networks to solve various real life problems is well understood. This paper proposes a mechanism for automated neural network based transport system with learning (MANTLE), an automated transport system with considerable advantages over previous attempts. The system uses a multilayer feed-forward neural network with back propagation learning. In addition, the design of MANTLE involves the convergence of a plethora of technologies like Global Positioning System (GPS), a geographic information system (GIS), and laser ranging. MANTLE can guide a mobile agent through a hostile and unfamiliar domain after being trained by a human user with domain expertise. One of the many areas in which MANTLE scores against the competition is that the system is completely domain independent and incurs substantially less processor overhead. MANTLE thus provides more functionality, even though it requires a lot less input as compared to other attempts in this field. This reduction in the size of the input vector translates into more efficient and faster processing. Another of MANTLE's hallmark features is its ability to negotiate turns and implement lane-changing maneuvers with a view to overtaking obstacles. It does this by employing a novel technique, selective net masking. A simulation of MANTLE's neural network was performed on a variety of network topologies, and the best network selected.

Computer-aided Prediction of the ZrO2 Nanoparticles' Effects on Tensile Strength and Percentage of Water Absorption of Concrete Specimens

The skill of experienced log analysts is emulated by a new computer technology called Simulated Neural Networks (SNNs). Two SNNs are applied to problems in log interpretation. The first SNN is applied to a trivial and conceptually familiar interpretation problem: learning to compute total porosity (PHIT) and apparent grain density (RHOMA) from log data and lithology inputs. The second SNN learned to identify lithofacies patterns as a response to input log data. Interesting and human-like results were obtained when these SNNs interpreted log data in a well which has been full bore cored from top to bottom. Both SNNs performed as well as experienced humans and at least as well or better than conventional sequential computer program interpretations of the same two analysis problems.

Analysis of Japanese logistics development mode based on neural network and computer simulation is conducted in this research. The Japanese government provides strong financial support for green logistics and implements various green environmental protection tax preferential policies. For then referring this advanced feature, the listed aspects are analyzed. (1) During the training and learning process, the state of the neuron nodes in each layer only affects the signals of the neurons in the lower layer, which are gradually transmitted through multiple layers. (2) The simulation with the clustering model is analyzed. The simulation results are combined with efficient information analysis and mining. The experimental results compared with the other have proven the suggestions are reasonable.

The major bottleneck in simulation of large-scale neural networks is the communication problem due to one-to-many neuron connectivity. Network-on-Chip concept has been proposed to address the problem. This work explores the drawback that is introduced by interconnection networks - a delay jitter. The preliminary experiment is held in the spiking neural network simulator introducing variable communicational delay to the simulation. The performance degradation is reported.

The field of neural networks is being investigated by many researchers in order to provide solutions to difficult problems in the area of manufacturing systems. Computer simulation of neural networks is an important part of this investigation. This paper applies concepts from an important trend in software engineering research, namely object-oriented programming, to model neural networks.

High-performance fault simulation is one of the essential and preliminary tasks in the process of online and offline testing of machine learning (ML) hardware. Deep neural networks (DNN), as one of the essential parts of ML programs, are widely used in many critical and non-critical applications in Systems-on-Chip and ASIC designs. Through fault simulation for DNNs, by increasing the number of neurons, the fault simulation time increases exponentially. However, the software architecture of neural networks and the lack of dependency between neurons in each inference layer provide significant opportunity for parallelism of the fault simulation time in a multi-processor platform. In this paper, a multi-thread technique for hierarchical fault simulation of neural network is proposed, targeting both permanent and transient faults. During the process of fault simulation the neurons for each inference layer will be distributed among the executing threads. Since in the process of hierarchical fault simulation, the faulty neuron demands proportionally enormous computation comparing to behavioural model of non-faulty neurons, the faulty neuron will be assigned to one thread while the rest of the neurons will be divided among the remaining threads. Experimental results confirm the time efficiency of the proposed fault simulation technique on multi-processor architectures.

Neuroimage: A Novel Highly Efficient Tool for Image Processing of in vivo Neural Networks

Neural networks: An overview of early research, current frameworks and new challenges