Search

The adoption of Artificial Neural Networks (ANNs) in safety-related applications is often avoided because it is difficult to rule out possible misbehaviors with traditional analytical or probabilistic techniques. In this paper we present N eV er, our tool for checking safety of ANNs. N eV er encodes the problem of verifying safety of ANNs into the problem of satisfying corresponding Boolean combinations of linear arithmetic constraints. We describe the main verification algorithm and the structure of N eV er. We present also empirical results confirming the effectiveness of N eV er on realistic case studies. [ABSTRACT FROM AUTHOR]

Abstract: Accurate prediction for the synthesis characteristics of hydraulic valve in industrial production plays an important role in decreasing the repair rate and the reject rate of the product. Recently, Support Vector Machine (SVM) as a highly effective mean of system modeling has been widely used for predicting. However, the important problem is how to choose the reasonable input parameters for SVM. In this paper, a hybrid prediction method (SA–SVM for short) is proposed by using simulated annealing (SA) and SVM to predict synthesis characteristics of the hydraulic valve, where SA is used to optimize the input parameters of SVM based prediction model. To validate the proposed prediction method, a specific hydraulic valve production is selected as a case study. The prediction results show that the proposed prediction method is applicable to forecast the synthesis characteristics of hydraulic valve and with higher accuracy. Comparing with Adaptive Neuro-Fuzzy Inference System (ANFIS) and Artificial Neural Networks (ANN) are also made. [Copyright &y& Elsevier]

This paper develops and applies neural network (NN) models to forecast regional employment patterns in Germany. Computer-aided optimization tools that imitate natural biological evolution to find the solution that best fits the given case (namely, genetic algorithms, GAs) are also used to detect the best NN structure. GA techniques are compared with more `traditional' techniques which require the supervision of experienced analysts. We test the performance of these techniques on a panel of 439 districts in West and East Germany. Since the West and East datasets have different time spans, the models are estimated separately for West and East Germany. The results show that the West and East NN models perform with different degrees of precision, mainly because of the different time spans of the two datasets. Automatic techniques for the choice of the NN architecture do not seem to outperform selection procedures based on the supervision of expert analysts. [ABSTRACT FROM AUTHOR]

Abstract: The facility layout problem is concerned with finding feasible locations for a set of interrelated objects that meet all design requirements and maximize design quality in terms of design preferences. The contribution of this paper is a new framework, named annealed neural network, for efficiently finding competitive solutions for the facility layout problem. This framework arises from the combination of Hopfield neural networks and simulated annealing. The first is a representation model of the layout problem and the second is a search algorithm for finding the optimum or near optimum solutions. The annealed neural network combines characteristics of the simulated annealing algorithm and the Hopfield neural network. Annealed neural network exhibits the rapid convergence of the neural network, while preserving the solution quality afforded by simulated annealing. Strategies for setting reasonable penalty factor in objective function and temperature in simulated annealing procedure were proposed. A case study of a hospital building with 28 facilities was employed to demonstrate that this model is rather efficient to solve the architectural layout problem, and it is amenable to fast computation for large layout problems. [Copyright &y& Elsevier]

The article presents information on a case study of forecasting nonlinear time series with feed-forward neural networks. The forecasting capabilities of feed-forward neural network (FFNN) models are compared to those of other competing time series models by carrying out forecasting experiments. In spite of the diverse applicability of neural networks in many different areas, much controversy surrounds their employment for tackling problems that can also be studied using well-established statistical models. One such controversial domain is time series forecasting. Following a description of FFNN models in the next section, an overview is given about the use of neural networks in time series prediction. Methods for rigorously comparing forecasts between FFNN models and statistical models are required for properly carrying out forecasting experiments. In terms of modelling speed, FFNN models work well. There are sensible approaches to construct FFNN models but, in fact, there are not many restrictions for building FFNN models because of their black box characteristics.

This article focuses on variable selection and mobility for identification of neural model. Neural networks are essentially devices for non-parametric statistical inference. This paper describes a methodology for neural model identification which facilitates hypothesis testing both for model adequacy and variable significance. The objective of a model identification procedure is to construct a well-specified estimator of the unknown function. A well-specified estimator will be a faithful representation of the regression, all the parameters and variables used by the estimator will be significant and the expectation of zero. In theory, it may be possible to devise a model-identification procedure which attempts to satisfy all these goals simultaneously. In practice, however, it is much easier to separate these goals into three distinct tasks model selection, variable significance testing, and model adequacy testing. The individual components of the methodology were evaluated in controlled simulations with artificial data. A case study in tactical asset allocation was used to concept-test the overall methodology in a non-trivial financial problem. Neural models were contrasted to multiple linear regression on the basis of model identification.

In recent years, artificial intelligence techniques such as artificial neural networks (ANN) and Support Vector Regression (SVM) have been well documented in ecological sciences. These methods can perfectly model complex and nonlinear structures, as well as with high processing power and quick computations in ecological sciences. Research on ecological issues with artificial intelligence methods can be useful and provided that the details of the use of these methods are necessary for readers. In this discussion, the discusser has tried to clarify the process of the paper of "Design of an integrated climatic assessment indicator (ICAI) for wheat production: A case study in Jiangsu Province, China" (doi: 10.1016/j.ecolind.2019.01.059). The discusser would like to call attention to some important points, which may be taken into consideration by the authors and other potential researchers. [ABSTRACT FROM AUTHOR]