Dynamic attention-integrated neural network for session-based news recommendation

This paper deals with the adaptive nonlinear identification and trajectory tracking via dynamic multilayer neural network (NN) with different timescales. Two NN identifiers are proposed for nonlinear systems identification via dynamic NNs with different timescales including both fast and slow phenomenon. The first NN identifier uses the output signals from the actual system for the system identification. In the second NN identifier, all the output signals from nonlinear system are replaced with the state variables of the NNs. The online identification algorithms for both NN identifier parameters are proposed using Lyapunov function and singularly perturbed techniques. With the identified NN models, two indirect adaptive NN controllers for the nonlinear systems containing slow and fast dynamic processes are developed. For both developed adaptive NN controllers, the trajectory errors are analyzed and the stability of the systems is proved. Simulation results show that the controller based on the second identifier has better performance than that of the first identifier.

One of the difficulties encountered in the application of reinforcement learning methods to real-world problem is the generalization of large-scale or continuous state space. In order to solve the curse of dimensionality problem caused by discretizing continuous state space, a kind of Q-learning method for continuous state space based on a dynamic Elman neural network was proposed in this paper. The inputs and the output of Elman network are the system state-action pair and the corresponding Q-value. That is, Elman network is used to estimate the Q-value of state-action pair on-line. Eligibility trace for connecting weights is introduced by borrowing ideas from the eligibility trace mechanism of state in temporal difference algorithm to improve the learning speed of neural network. Computer simulations on mountain car control illustrate the performance and applicability of the proposed Q-learning scheme.

An increasing number of applications of dynamic neural networks has been developed for digital signal processing (DSP), dynamic neural networks are feedforward neural networks with commonly used scalar synapses replaced by linear filters. This provides feedforward neural networks with the capability of performing dynamic mappings, which depend on past input values, dynamic neural networks are suitable for time series prediction, nonlinear system identification, and signal processing applications. Their most popular types are Finite Impulse Response (FIR) neural networks, which are obtained by replacing synapses with finite impulse response filters. Due to their guaranteed stability characteristic and easy to minimize error surface they have been used with great success in many applications such as signal enhancement, noise cancellation, classification of input patterns, system identification, prediction, and control.. Most of the works on system identification using neural networks are based on multilayer feedforward neural networks with backpropagation learning or more efficient variations of this algorithm, an elegant method for training layered networks. This paper is based on work in a Dynamic System Modeling (DYSMO) and as an application for speed control of DC motor drive.

Several mathematical approaches to studying analytically the dynamics of neural networks rely on mean-field approximations, which are rigorously applicable only to networks of infinite size. However, all existing real biological networks have finite size, and many of them, such as microscopic circuits in invertebrates, are composed only of a few tens of neurons. Thus, it is important to be able to extend to small-size networks our ability to study analytically neural dynamics. Analytical solutions of the dynamics of small-size neural networks have remained elusive for many decades, because the powerful methods of statistical analysis, such as the central limit theorem and the law of large numbers, do not apply to small networks. In this article, we critically review recent progress on the study of the dynamics of small networks composed of binary neurons. In particular, we review the mathematical techniques we developed for studying the bifurcations of the network dynamics, the dualism between neural activity and membrane potentials, cross-neuron correlations, and pattern storage in stochastic networks. Then, we compare our results with existing mathematical techniques for studying networks composed of a finite number of neurons. Finally, we highlight key challenges that remain open, future directions for further progress, and possible implications of our results for neuroscience.

The article is devoted to the problem of optimization of search request ranking algorithms in the digital information retrieval system. The algorithm of functioning of the neural network ranking unit based on Hopfield neural network is built. The ability to generate a ranked list of pages found as a result of the request in the digital information retrieval system can be provided by solving two problems of integer optimization: the problem of assignment of combinatorial sets of criteria for assessing the relevance of web page search and the problem of sorting of numbers—relevance values. The architecture of the neural network model based on the dynamic Hopfield neural network with binary output function designed for combinatorial optimization of the final list of documents found in the digital information retrieval system was synthesized. Promising variants of neural network models with binary output function of neurons for synthesis of the optimal evaluation plan with a combinatorial set of criteria by solving the problem of assignment were built. It has been proven that the built models differ in the rules for determining the coefficients of synaptic connections and external shifts; each of the created rules can be used independently or in different combinations with one another. In the course of analytical research, it was found that the optimization formulation of the problem of sorting of relevance values of search pages is identical to the problem of assignment of combinatorial groups of evaluation criteria provided that the elements of the performance matrix of the latter are defined as linear combinations of relevance values.

Recently, neural networks have been increasingly applied to remote sensing imagery classification. The conventional neural network classifier performs learning from the representative information within a problem domain on a one-pixel-one-class basis; therefore, class mixture and the degree of membership of a pixel are generally not taken into account, often resulting in a poor classification accuracy. Based on the framework of a dynamic learning neural network (DL), this communications proposes a fuzzy version (FDL) based on two steps: network representation of fuzzy logic and assignment of membership. Comparisons between the DL and FDL are made by applying both neural networks to SAR image classification. Experimental results show that the FDL has faster convergence rate than that of DL. In addition, the separability between similar classes is improved. Moreover, the classification results match better with ground truth.

Intelligent Multivariable Modeling of Blast Furnace Molten Iron Quality Based on Dynamic AGA-ANN and PCA

Neural networks are efficient and effective in predicting system states in hydrology. However, most current approaches lack hydrological flow partitioning, do not allow for training on measurements of multiple variables, or lack capability to tightly integrate physically-based components. To address these shortcomings I propose and evaluate an approach referred to below as Dynamical System Neural Network (DSNN). DSNN is a feedforward neural network with an architecture that resembles the organisation in components of the real-world system it represents. In hydrology, the DSNN represents each water flow (e.g. seepage, snow melt) by a collection of input, hidden, and output neural layers, where each input is the state of a hydrological storage (e.g. groundwater storage influencing seepage) or other variable (e.g. air temperature influencing snow melt). These components are interconnected to form a single neural network of the complete dynamical system considered, where all storages and flows are explicitly quantified. If physical understanding of a flow and its parameterization is available, a known formulation can be used as a replacement of a neural network component. The DSNN is applied forward in time, backpropagating gradients over all timesteps. It can be run in spatially lumped or semi-distributed mode. To demonstrate the approach, a DSNN is presented of the Austrian Dorfertal (Kals) Alpine catchment containing snow and subsurface water storages and associated flows including streamflow. The DSNN is trained, validated, and tested on daily streamflow over ~40 years. To explore the capability of the DSNN in estimating the magnitude and dynamics of internal system storages (snow water equivalent, subsurface water storage) and flows (evapotranspiration, sublimation, snowmelt, seepage), the DSNN is first trained and tested with streamflow data generated by a conceptual model. The DSNN turns out to be capable of reproducing - with a satisfactory level of precision - the system states and fluxes calculated by the conceptual model, with decreasing performance when measurement error is added to the artificially generated streamflow data before training. To explore its predictive performance, the DSNN is applied on measured streamflow data for the Dorfertal, comparing multiple DSNN setups that represent all flows as neural network components or only a subset of flows where remaining flows are represented with a standard conceptual model (e.g. linear reservoir). Preliminary results indicate that in predictive performance, in most setups, the DSNN outperforms a standard conceptual model trained on the same streamflow data, with NSE values for testing of 0.74 and 0.71, respectively. This preliminary result indicates DSNN to be a promising approach for blending process-based and neural network based modelling as well as for training (i.e. calibration) of neural network models on measurements of multiple hydrological variables as these are all explicitly represented by the DSNN and can thus be incorporated in the loss function (e.g. streamflow, snow depth, groundwater, evapotranspiration).

In many applications neural networks must process or generate time series, and various network paradigms exist for this purpose. Two prominent examples are time-delay neural networks (TDNN), which are known for their noise suppression capability, and NARX (nonlinear autoregressive models with exogenous inputs) networks, which have a powerful modeling ability (at least Turing equivalence). In this article, we suggest a combination of these two approaches, called dynamic neural network (DYNN), which unifies the particular advantages. Efficient training algorithms are needed to adjust the weights of DYNN. Here, we describe an algorithm for the computation of first-order information about the error surface: temporal backpropagation through time (TBPTT). Essentially, this algorithm is a combination of temporal backpropagation (used for TDNN) and backpropagation through time (used for NARX). The first-order information is then utilized to apply the scaled conjugate gradient (SCG) learning algorithm which approximates second-order with first-order information. The benefits of this approach are shown by means of two benchmark data sets: logistic map and building. It is shown that SCG for DYNN is significantly faster and more accurate than other learning algorithms (e.g. TBPTT, resilient propagation, memoryless Quasi-Newton).

Given e-commerce scenarios that user profiles are invisible, session-based recommendation is proposed to generate recommendation results from short sessions. Previous work only considers the user's sequential behavior in the current session, whereas the user's main purpose in the current session is not emphasized. In this paper, we propose a novel neural networks framework, i.e., Neural Attentive Recommendation Machine (NARM), to tackle this problem. Specifically, we explore a hybrid encoder with an attention mechanism to model the user's sequential behavior and capture the user's main purpose in the current session, which are combined as a unified session representation later. We then compute the recommendation scores for each candidate item with a bi-linear matching scheme based on this unified session representation. We train NARM by jointly learning the item and session representations as well as their matchings. We carried out extensive experiments on two benchmark datasets. Our experimental results show that NARM outperforms state-of-the-art baselines on both datasets. Furthermore, we also find that NARM achieves a significant improvement on long sessions, which demonstrates its advantages in modeling the user's sequential behavior and main purpose simultaneously.

Recurrent neural networks unlike feed-forward networks are able to process inputs with time context. The key role in this process is played by the dynamics of the network, which transforms input data to the recurrent layer states. Several authors have described and analyzed dynamics of small sized recurrent neural networks with two or three hidden units. In our work we introduce techniques that allow to visualize and analyze the dynamics of large recurrent neural networks with dozens units, reveal both stable and unstable points (attractors and saddle points), which are important to understand the principles of successful task processing. As a practical example of this approach, dynamics of the simple recurrent network trained by two different training algorithms on context-free language anbnwas studied.

This paper investigates the use of static and dynamic neural networks in phoneme recognition. Besides this, the paper also proposes a cooperative static and dynamic neural networks model. The cooperative model integrates a decision system for phoneme recognition. Mel cepstrum coding has been applied to represent speech signal in frames. Features from the selected frames are used to train neural networks based models. The comparative study show that the proposed cooperative model provides more accurate recognition rates both in auto-coherence test and generalization test.

Power battery is a core energy system of electric vehicle (EV). The accurate state of charge (SOC) prediction is the most basic and important for the battery management system. There are many shortcomings in the traditional method of predicting SOC, such as the accumulation of errors with time and the difficulty of on-line prediction, etc. So a Dynamic Neural Network by Kalman filter (KDNN) is proposed to predict the SOC of electric vehicles. The weights of dynamic neural network is updated by KF. At the same time, in order to reduce the randomness of the prediction algorithm, the Genetic algorithm (GA) is used to optimize the initial weights of KDNN. Then the battery data from the experiment during the car driving are used to test the performance of KDNN with GA. And the comparisons are done between the new KDNN and BP prediction methods. The results demonstrate that the new KDNN with GA prediction modeling method has obvious superiority compared with BP neural network.

A spatial–temporal graph neural network framework for automated software bug triaging

This paper proposes a dynamic neural network (DNN) to solve the time-varying shortest path problem with hop-constraint (HC-TSPP). The purpose of HC-TSPP is to find a path with the shortest transmission time and the restricted number of arcs. The proposed DNN is a novel neural network based on dynamic neurons. All neurons on DNN are computing in parallel, and each dynamic neuron is composed of seven parts: input, wave receiver, filter, status memorizer, wave generator, wave sender, and output. Wave is the carrier of neuron communication, and each wave is composed of three parts. The shortest path report is based on the first wave that reaches the destination node and satisfies the hop constraint. The example and experimental results based on the Internet dataset show that the proposed algorithm can arrive at the global optimal solution and outperform the existing algorithm (viz. Dijkstra algorithm).