Two artificial neural networks (ANN), back-propagation neural network (BPNN) and the radial basis function neural network (RBFNN), are proposed to predict the carbonation depth of prestressed concrete. In order to generate the training and testing data for the ANNs, an accelerated carbonation experiment was carried out, and the influence of stress level of concrete on carbonation process was taken into account especially. Then, based on the experimental results, the BPNN and RBFNN models which all take the stress level of concrete, water-cement ratio, cement-fine aggregate, cement-coarse aggregate ratio and testing age as input parameters were built and all the training and testing work was performed in MATLAB. It can be found that the two ANN models seem to have a high prediction and generalization capability in evaluation of carbonation depth, and the largest absolute percentage errors of BPNN and RBFNN are 10.88% and 8.46%, respectively. The RBFNN model shows a better prediction precision in comparison to BPNN model.

Chunhua Lu

Ronggui Liu

Predicting Carbonation Depth of Prestressed Concrete under Different Stress States Using Artificial Neural Network

Artificial neural networks (ANNs) are a form of artificial intelligence that has proved to provide a high level of competency in solving many complex engineering problems that are beyond the computational capability of classical mathematics and traditional procedures. In particular, ANNs have been applied successfully to almost all aspects of geotechnical engineering problems. Despite the increasing number and diversity of ANN applications in geotechnical engineering, the contents of reported applications indicate that the progress in ANN development and procedures is marginal and not moving forward since the mid-1990s. This paper presents a brief overview of ANN applications in geotechnical engineering, briefly provides an overview of the operation of ANN modeling, investigates the current research directions of ANNs in geotechnical engineering, and discusses some ANN modeling issues that need further attention in the future, including model robustness; transparency and knowledge extraction; extrapolation; uncertainty.

Mohamed A. Shahin

Mark B. Jaksa

Holger R. Maier

Recent Advances and Future Challenges for Artificial Neural Systems in Geotechnical Engineering Applications

The bioelectric potentials associated with muscle activity constitute the electromyogram (EMG). These EMG signals are low-frequency and lower-magnitude signals. In this paper, it is presented that Jordan/Elman neural network can be effectively used for EMG signal noise removal, which is a typical nonlinear multivariable regression problem, as compared with other types of neural networks. Different neural network (NN) models with varying parameters were considered for the design of adaptive neural-network-based filter which is a typical SISO system. The performance parameters, that is, MSE, correlation coefficient, <i>N/P</i>, and <i>t</i>, are found to be in the expected range of values.

V. R. Mankar

A. A. Ghatol

Design of Adaptive Filter Using Jordan/Elman Neural Network in a Typical EMG Signal Noise Removal

In order to simulate neuronal electrical activities, we must estimate the dynamics of channel conductances from physiological experimental data. However, this approach requires the formulation of differential equations that express the time course of channel conductance. On the other hand, if the dynamics are automatically estimated, neuronal activities can be easily simulated. By using a recurrent neural network (RNN), it is possible to estimate the dynamics of channel conductances without formulating the differential equations. In the present study, we estimated the dynamics of the <svg style="vertical-align:-0.16379pt;width:27.387501px;" id="M1" height="14.3125" version="1.1" viewBox="0 0 27.387501 14.3125" width="27.387501" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg"><g transform="matrix(.017,-0,0,-.017,.062,14.05)"><path id="x4E" d="M719 650v-28q-43 -2 -62 -15t-22 -44q-6 -47 -6 -169v-403h-31l-426 524h-2v-251q0 -111 6 -169q4 -37 24 -50.5t72 -16.5v-28h-237v28q45 2 64.5 16t23.5 49q6 62 6 171v220q0 54 -3 68.5t-17 32.5q-16 19 -34.5 26.5t-54.5 10.5v28h147l418 -502h3v246q0 117 -7 166&#xA;q-4 34 -24.5 47t-73.5 15v28h236z" /></g><g transform="matrix(.017,-0,0,-.017,12.251,14.05)"><path id="x61" d="M433 39l-85 -51q-27 0 -47 20q-16 16 -24 44q-5 -3 -24 -16.5t-27 -18.5t-23 -13.5t-25 -12t-19 -3.5q-52 0 -86 36.5t-34 85.5q0 71 81 99q128 43 155 65v17q0 54 -23 83.5t-62 29.5q-28 0 -45 -19t-29 -66q-7 -23 -29 -23q-15 0 -29 13t-14 30t31 40q25 19 68.5 40.5&#xA;t81.5 29.5q49 0 82 -27q45 -39 45 -123v-185q0 -60 35 -60q18 0 36 11zM275 84v156q-12 -6 -49 -23t-41 -19q-31 -14 -46 -31t-15 -43q0 -35 22.5 -55.5t48.5 -20.5q19 0 42 10.5t38 25.5z" /></g><g transform="matrix(.012,-0,0,-.012,19.713,5.887)"><path id="x2B" d="M535 230h-212v-233h-58v233h-213v50h213v210h58v-210h212v-50z" /></g></svg> and <svg style="vertical-align:-0.0pt;width:19.174999px;" id="M2" height="14.1125" version="1.1" viewBox="0 0 19.174999 14.1125" width="19.174999" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg"><g transform="matrix(.017,-0,0,-.017,.062,14.05)"><path id="x4B" d="M667 0h-153l-248 291q-28 32 -54 27v-190q0 -65 14.5 -80t77.5 -20v-28h-260v28q58 5 71 20t13 80v392q0 48 -6 66.5t-22 25.5t-59 10v28h258v-28q-59 -5 -73 -20.5t-14 -81.5v-175q29 6 70 45q101 97 167 178q23 27 20 37t-36 14l-23 3v28h240v-28q-57 -5 -86 -18.5&#xA;t-76 -56.5q-40 -35 -177 -175q97 -124 235 -272q39 -41 61.5 -54t59.5 -18v-28z" /></g><g transform="matrix(.012,-0,0,-.012,11.5,5.887)"><use xlink:href="#x2B" /></g></svg> conductances of a squid giant axon using two different fully connected RNNs and were able to reproduce various neuronal activities of the axon. The reproduced activities were an action potential, a threshold, a refractory phenomenon, a rebound action potential, and periodic action potentials with a constant stimulation. RNNs can be trained using channels other than the <svg style="vertical-align:-0.16379pt;width:27.387501px;" id="M3" height="14.3125" version="1.1" viewBox="0 0 27.387501 14.3125" width="27.387501" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg"><g transform="matrix(.017,-0,0,-.017,.062,14.05)"><use xlink:href="#x4E" /></g><g transform="matrix(.017,-0,0,-.017,12.251,14.05)"><use xlink:href="#x61" /></g><g transform="matrix(.012,-0,0,-.012,19.713,5.887)"><use xlink:href="#x2B" /></g></svg> and <svg style="vertical-align:-0.0pt;width:19.174999px;" id="M4" height="14.1125" version="1.1" viewBox="0 0 19.174999 14.1125" width="19.174999" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns="http://www.w3.org/2000/svg"><g transform="matrix(.017,-0,0,-.017,.062,14.05)"><use xlink:href="#x4B" /></g><g transform="matrix(.012,-0,0,-.012,11.5,5.887)"><use xlink:href="#x2B" /></g></svg> channels. Therefore, using our RNN estimation method, the dynamics of channel conductance can be automatically estimated and the neuronal activities can be simulated using the channel RNNs. An RNN can be a useful tool to estimate the dynamics of the channel conductance of a neuron, and by using the method presented here, it is possible to simulate neuronal activities more easily than by using the previous methods.

Masaaki Takahashi

Kiyohisa Natsume

Automatic Estimation of the Dynamics of Channel Conductance Using a Recurrent Neural Network

Reservoir computing (RC) is a new framework for neural computation. A reservoir is usually a recurrent neural network with fixed random connections. In this article, we propose an RC model in which the connections in the reservoir are modifiable. Specifically, we consider correlation-based learning (CBL), which modifies the connection weight between a given pair of neurons according to the correlation in their activities. We demonstrate that CBL enables the reservoir to reproduce almost the same spatiotemporal activity patterns in response to an identical input stimulus in the presence of noise. This result suggests that CBL enhances the robustness in the generation of the spatiotemporal activity pattern against noise in input signals. We apply our RC model to trace eyeblink conditioning. The reservoir bridged the gap of an interstimulus interval between the conditioned and unconditioned stimuli, and a readout neuron was able to learn and express the timed conditioned response.

Tadashi Yamazaki

Shigeru Tanaka

Robust Reservoir Generation by Correlation-Based Learning

The present paper proposes a recurrent neural network model and learning algorithm that can acquire the ability to generate desired multiple sequences. The network model is a dynamical system in which the transition function is a contraction mapping, and the learning algorithm is based on the gradient descent method. We show a numerical simulation in which a recurrent neural network obtains a multiple periodic attractor consisting of five Lissajous curves, or a Van der Pol oscillator with twelve different parameters. The present analysis clarifies that the model contains many stable regions as attractors, and multiple time series can be embedded into these regions by using the present learning method.

Jun Namikawa

Jun Tani

Building Recurrent Neural Networks to Implement Multiple Attractor Dynamics Using the Gradient Descent Method

About this journal

Journal profile

Meet the editorial board

Editor spotlight

Submit your proposal

Do you think there is an emerging area of research that really needs to be highlighted? Or an existing research area that has been overlooked or would benefit from deeper investigation? Raise the profile of a research area by leading a Special Issue.

Advances in Artificial Neural Systems