Adaptive Learning Algorithm Research Articles

Robust adaptive filter for shipborne kinematic positioning and velocity determination during the Baltic Sea experiment

Information on trajectory and attitude is essential for analyzing gravimetric data collected on kinematic platforms. Usually, a Kalman filter is used to obtain high-accuracy positional and velocity information. However, this can be affected by measurement outliers and by state disturbances that occur frequently under a fast-changing environment. To overcome these problems, a robust adaptive Kalman filtering algorithm is applied for state estimates, which introduces an equivalent weight to resist measurement outliers and an optimal adaptive factor to balance the contributions of the kinematic model information and the measurements. In addition to the conventional robust estimator, an improved Current Statistical (CS) model is proposed. The improved CS model adopts a variance adaptive learning algorithm, and it can perform self-adaptation of acceleration variance with the innovation information; thus, it can overcome the shortcoming of lower tracking accuracy and avoid setting the maximum acceleration. Following a gravimetry campaign on the Baltic Sea, it is shown in theory and in practice that the robust adaptive Kalman filter is not only simple in its calculation but also more reliable in controlling the colored observation noise and kinematic state disturbance compared with the classical Kalman filter. The improved CS model performs best, especially when analyzing the positioning errors at the turns due to the target maneuvering. Compared to the CS model, the RMS values of the positional estimates derived from the improved CS model decrease by almost 30% in the horizontal direction, and no significant improvement in the vertical direction is found.

What does the yield curve imply about investor expectations?

We use daily data to model investors’ expectations of U.S. yields, at different maturities and forecast horizons. We consider two adaptive learning algorithms to characterize the conditional yield forecasts. Our framework yields the first empirical estimates of the pace of learning by investors. The superior performance of the endogenous learning mechanism suggests that investors account for structural change and respond to significant, persistent deviations by modifying the amount of information they use. Our results provide strong empirical motivation to use the class of adaptive learning models considered here for modeling and analyzing expectation formation by investors.

A novel adaptive learning algorithm for low-dimensional feature space using memristor-crossbar implementation and on-chip training

Proposing an efficient algorithm with an appropriate hardware implementation has always been an interesting and a rather challenging field of research in Artificial Intelligence (AI). Fuzzy logic is one of the techniques that can be used for accurate and high-speed modeling as well as controlling complex and nonlinear systems. The “defuzzification” process during the test phase as well as the repetitive processes in order to find the optimal parameters during the training phase, lead to some serious limitations in real-time applications and hardware implementation of these algorithms. The proposed algorithm employs Ink Drop Spread (IDS) concept to mimic the functionality of human brain. In this algorithm, learning is based on the distance between training data and the “learning plane”. Unlike previous algorithms, the new one does not need to partition nor the input space neither the calculation of IDS plane features. Besides, the output is obtained without using the optimization methods. The proposed algorithm is a numerical foundation that does not encounter a processing problem and lack of memory in dealing with different datasets consisting of a large number of samples. This algorithm can be efficiently implemented on memristor crossbar/CMOS hardware platform in terms of area and speed. This hardware has the ability to learn and adapt to the environment regardless of a host system (on-chip learning capability). Finally, to verify the performance of the proposed algorithm, it has been compared to ALM, RBF and PNN algorithms which have a similar functionality.

Reservoir of diverse adaptive learners and stacking fast hoeffding drift detection methods for evolving data streams

The last decade has seen a surge of interest in adaptive learning algorithms for data stream classification, with applications ranging from predicting ozone level peaks, learning stock market indicators, to detecting computer security violations. In addition, a number of methods have been developed to detect concept drifts in these streams. Consider a scenario where we have a number of classifiers with diverse learning styles and different drift detectors. Intuitively, the current 'best' (classifier, detector) pair is application dependent and may change as a result of the stream evolution. Our research builds on this observation. We introduce the $\mbox{Tornado}$ framework that implements a reservoir of diverse classifiers, together with a variety of drift detection algorithms. In our framework, all (classifier, detector) pairs proceed, in parallel, to construct models against the evolving data streams. At any point in time, we select the pair which currently yields the best performance. We further incorporate two novel stacking-based drift detection methods, namely the $\mbox{FHDDMS}$ and $\mbox{FHDDMS}_{add}$ approaches. The experimental evaluation confirms that the current 'best' (classifier, detector) pair is not only heavily dependent on the characteristics of the stream, but also that this selection evolves as the stream flows. Further, our $\mbox{FHDDMS}$ variants detect concept drifts accurately in a timely fashion while outperforming the state-of-the-art.

A Dual Stage Well Dressed Crawler using Adaptive Learning Algorithm for Collecting Deep - Web Interfaces

A Dual Stage Well Dressed Crawler using Adaptive Learning Algorithm for Collecting Deep - Web Interfaces

Нейромережеві засоби прогнозування споживання енергоресурсів

Досліджено та обґрунтовано вибір нейромережевих структур для оброблення статистичних даних з метою прогнозування та виявлення аномальних показників споживання енергоресурсів. Показано, що системам на основі нейронних мереж завжди протиставлялись експертні системи, які, на відміну від перших, очевидно програмувались. Середовище, в якому працює система, не завжди є статичним і потрібні методи опрацювання даних, які могли б адекватно реагувати на зміну середовища та вміти відповідно адаптувати отримувані результати. Нейронні мережі володіють такою особливістю, як вміння навчатись. Ця особливість і є основним аргументом для застосування таких структур у системах управління енергоефективністю. Розроблена архітектура мережі та застосований процес навчання дав змогу прогнозувати показники спожитої електроенергії з урахуванням багатьох параметрів. Особливістю розробленої архітектури є можливість здійснювати перенавчання у процесі функціонування, не перериваючи його. Використання адаптивного та безперервного навчання нейромережі дасть змогу виявляти аномальні показники даних. Точність такого виявлення було перевірено на реальній вибірці даних. Аналіз отриманих результатів показує, що використання нейронних мереж хоч і потребує швидкодії і часу на навчання, проте, під час класифікації вхідного вектора, швидкодія нейронної мережі перевищує будь-який алгоритм кластеризації.

Optimizing the learning rate for adaptive estimation of neural encoding models.

Closed-loop neurotechnologies often need to adaptively learn an encoding model that relates the neural activity to the brain state, and is used for brain state decoding. The speed and accuracy of adaptive learning algorithms are critically affected by the learning rate, which dictates how fast model parameters are updated based on new observations. Despite the importance of the learning rate, currently an analytical approach for its selection is largely lacking and existing signal processing methods vastly tune it empirically or heuristically. Here, we develop a novel analytical calibration algorithm for optimal selection of the learning rate in adaptive Bayesian filters. We formulate the problem through a fundamental trade-off that learning rate introduces between the steady-state error and the convergence time of the estimated model parameters. We derive explicit functions that predict the effect of learning rate on error and convergence time. Using these functions, our calibration algorithm can keep the steady-state parameter error covariance smaller than a desired upper-bound while minimizing the convergence time, or keep the convergence time faster than a desired value while minimizing the error. We derive the algorithm both for discrete-valued spikes modeled as point processes nonlinearly dependent on the brain state, and for continuous-valued neural recordings modeled as Gaussian processes linearly dependent on the brain state. Using extensive closed-loop simulations, we show that the analytical solution of the calibration algorithm accurately predicts the effect of learning rate on parameter error and convergence time. Moreover, the calibration algorithm allows for fast and accurate learning of the encoding model and for fast convergence of decoding to accurate performance. Finally, larger learning rates result in inaccurate encoding models and decoders, and smaller learning rates delay their convergence. The calibration algorithm provides a novel analytical approach to predictably achieve a desired level of error and convergence time in adaptive learning, with application to closed-loop neurotechnologies and other signal processing domains.

Fault Estimation and Accommodation for a Class of Nonlinear System Based on Neural Network Observer

The problem of fault estimation and accommodation of nonlinear systems with disturbances is studied using adaptive observer and neural network techniques. A robust adaptive learning algorithm based on switching βs-modification is developed to realize the accurate and fast estimation of unknown actuator faults or component faults. Then a fault tolerant controller is designed to restore system performance. Dynamic error convergence and system stability can be guaranteed by Lyapunov stability theory. Finally, simulation results of quadrotor helicopter attitude systems are presented to illustrate the efficiency of the proposed techniques.

Recurrent fuzzy wavelet neural networks based on robust adaptive sliding mode control for industrial robot manipulators

A robust adaptive control method is proposed in this paper based on recurrent fuzzy wavelet neural networks (RFWNNs) system for industrial robot manipulators (IRMs) to improve high accuracy of the tracking control. The RFWNNs consist of four layers, and second layer has the feedback connections. Wavelet basis function is used as fuzzy membership function. In general, it is not easy to adopt a model-based method to achieve this control object due to the uncertainties of the IRM, such as unknown dynamic, disturbances and parameter variations. To solve this problem, all the parameters of the RFWNNs system are tuned online by an adaptive learning algorithm, and online adaptive control laws are determined by Lyapunov stability theorem. In addition, the robust controller is designed to deal with the approximation error, optimal parameter vectors and higher-order terms in Taylor series. Therefore, with the proposed control, the desired tracking performance, stability and robustness of the closed-loop manipulators system are guaranteed. The simulations and experimental performed on a three-link IRMs are provided in comparison with fuzzy wavelet neural network and robust neural fuzzy network to demonstrate the effectiveness and robustness of the proposed RFWNNs methodology.

Research of Adaptive Neighborhood Incremental Principal Component Analysis and Locality Preserving Projection Manifold Learning Algorithm

In view of the incremental learning problem of manifold learning algorithm, an adaptive neighborhood incremental principal component analysis (PCA) and locality preserving projection (LPP) manifold learning algorithm is presented, and the incremental learning principle of algorithm is introduced. For incremental sample data, the adjacency and covariance matrices are incrementally updated by the existing samples; then the dimensionality reduction results of the incremental samples are estimated by the dimensionality reduction results of the existing samples; finally, the dimensionality reduction results of the incremental and existing samples are updated by subspace iteration method. The adaptive neighborhood incremental PCA-LPP manifold learning algorithm is applied to processing of gearbox fault signals. The dimensionality reduction results by incremental learning have very small error, compared with those by batch learning. Spatial aggregation of the incremental samples is basically stable, and fault identification rate is increased.

A novel tracking deep wavelet auto-encoder method for intelligent fault diagnosis of electric locomotive bearings

Abstract The condition monitoring of electric locomotive has attracted more and more attention due to its significance for improving the security, reliability and automation level. In this paper, a novel tracking deep wavelet auto-encoder (TDWAE) method is proposed for the intelligent fault diagnosis of electric locomotive bearings. Firstly, Gaussian wavelet function is adopted as the activation function to design wavelet auto-encoder (WAE), which can greatly enhance the quality of the features learned from the raw vibration signals of bearings. Secondly, a deep wavelet auto-encoder (DWAE) is constructed with several WAEs for higher-level feature learning and automatic fault diagnosis. Finally, an adaptive tracking learning algorithm is developed for flexibly determining the learning rate to further improve the diagnosis performance. The proposed method is applied to analyze the experimental vibration signals collected from electric locomotive bearings, and the results demonstrate that the proposed method is more effective than the traditional methods and standard deep auto-encoder.

Optimal guaranteed cost control of uncertain non‐linear systems using adaptive dynamic programming with concurrent learning

In this study, the authors study the optimal guaranteed cost control problem for a class of non-linear uncertain systems based on adaptive dynamic programming (ADP) with concurrent learning. A neural network-based approximate optimal guaranteed cost control design is developed not only to ensure the system stability for all admissible uncertainties but also to achieve a minimal guaranteed cost. First, the optimal guaranteed cost control problem is transformed into an optimal control problem of the nominal system by properly modifying the cost function to account for all possible uncertainties. Then based on ADP, an adaptive optimal learning algorithm is proposed for the nominal system by using a single critic network with concurrent learning to approximate the solution of Hamilton–Jacobi–Bellman equation. To relax the demands of the persistence of excitation condition, the recorded past data are used simultaneously with the current data for the adaptation of the critic network weights. Besides, an additional adjusting term is employed to stabilise the system and relax the requirement for an initial stabilising control. Moreover, uniform ultimate boundedness of the closed-loop system is guaranteed by using the Lyapunov approach. Finally, three simulation examples are provided to demonstrate the effectiveness of the proposed approach.

Music auto-tagging based on the unified latent semantic modeling

We proposed a music auto-tagging approach based on the latent space modeling both for music context and content. First, we introduce the latent semantic analysis for music tags with Sparse Nonnegative Matrix Factorization. Then the music contents semantics will be learnt by decomposing the music content into a pre-trained dictionary and an adaptive dictionary learning algorithm is proposed. Finally, the two latent spaces will be associated with a certain subspace mapping algorithm. The experimental results show that our proposed approach outperforms the state-of-the-art auto-tagging systems when applied to the CAL500 dataset in the 5-fold cross-validation experiments.

Adaptive Fuzzy Fractional-Order Nonsingular Terminal Sliding Mode Control for a Class of Second-Order Nonlinear Systems

This paper investigates a novel adaptive fuzzy fractional-order nonsingular terminal sliding mode controller (AFFO-NTSMC) for second-order nonlinear dynamic systems. The technique of fractional calculus and nonsingular terminal sliding mode control (NTSMC) are combined to establish fractional-order NTSMC (FO-NTSMC), in which a new fractional-order (FO) nonsingular terminal sliding mode (NTSM) surface is proposed. Then, a corresponding controller is designed to provide robustness, high performance control, finite time convergence in the presence of uncertainties and external disturbances. Furthermore, a fuzzy system with online adaptive learning algorithm is derived to eliminate the chattering phenomenon in conventional sliding mode control (SMC). The stability of the closed-loop system is rigorously proven. Numerical simulation results are presented to demonstrate the effectiveness of the proposed control method.

Adaptive Learning Algorithms for Simulation-Based Dynamic Traffic User Equilibrium

We propose two classes of algorithms for achieving the user equilibrium in simulation-based dynamic traffic assignment with special attention given to the interactions between travel information and route choice behavior. A driver is assumed to perform day-to-day route choice repeatedly and experience payoffs with unknown noise. The driver adaptively changes his/her action in accordance with the payoffs that are initially unknown and must be estimated over time due to noisy observations. To solve this problem, we develop a multi-agent version of Q-learning to estimate the payoff functions using novel forms of the e−greedy learning policy. We apply this Q-learning scheme for the simulation-based DTA in which traffic flow and travel times of routes in the traffic network are generated by a microscopic traffic simulator based on cellular automaton. Finally, we provide simulation examples to show convergence of our algorithms to Nash equilibrium and effectiveness of the best-route provision services.

Integrated soft sensor with wavelet neural network and adaptive weighted fusion for water quality estimation in wastewater treatment process

It is difficult to estimate the water quality of the wastewater treatment process, because the operating conditions are frequently changed. This paper gives an effective adaptive estimation method, which uses Hammerstein with wavelet neural networks, adaptive weighted fusion, and approximate linear dependence (ALD) analysis. Adaptive stable learning algorithm for the local Hammerstein with wavelet neural networks is proposed. A novel synchronous learning of fusion weighs is discussed. On-line calibration of operating centers with ALD improves the estimation accuracy. The experimental results show that the proposed estimation method for the water quality COD (Chemical Oxygen Demand) is satisfied compared with the laboratory results even when the operating conditions are changed frequently.

Sampling Adaptive Learning Algorithm for Mobile Blind Source Separation

Learning rate plays an important role in separating a set of mixed signals through the training of an unmixing matrix, to recover an approximation of the source signals in blind source separation (BSS). To improve the algorithm in speed and exactness, a sampling adaptive learning algorithm is proposed to calculate the adaptive learning rate in a sampling way. The connection for the sampled optimal points is described through a smoothing equation. The simulation result shows that the performance of the proposed algorithm has similar Mean Square Error (MSE) to that of adaptive learning algorithm but is less time consuming.

A Cognitive Vision Method for Insect Pest Image Segmentation

Segmentation of pests from crop leaves is the prerequisite step for pest intelligent diagnosis. To improve the accuracy and stability of segmentation, a cognitive segmentation approach to pest images is present in this paper. The method works in the follow way. First, a pest image is divided into blocks via an image block processing method. Second, an adaptive learning algorithm is used to accurately select the initial cluster centers. Third, preliminary segmentation results are achieved using K-means clustering. Finally, three digital morphological features of an ellipse are adopted to remove leaf veins. Segmentation experiments were performed on crop images of whiteflies. Compared with the conventional methods, the proposed cognitive segmentation method was accurate and robust for segmentation of whitefly images, and will provide a foundation for further identifying these pests.

Multi-agent reinforcement learning-based approach for controlling signals through adaptation

In this paper, we present a multi-agent reinforcement learning-based approach for controlling traffic signals. The aim is to use a multi-agent system with learning abilities for controlling and optimising traffic lights. We consider in this study the Q-learning algorithm, where the states are computed from average queue length in approaching links. The action space is modelled offline by using different time splits. The adaptation of the considered learning optimal policy through online learning is introduced to deal with the change of the environment. The simulation results show the effectiveness of the proposed adaptive learning algorithm.

Multi-agent reinforcement learning-based approach for controlling signals through adaptation

In this paper, we present a multi-agent reinforcement learning-based approach for controlling traffic signals. The aim is to use a multi-agent system with learning abilities for controlling and optimising traffic lights. We consider in this study the Q-learning algorithm, where the states are computed from average queue length in approaching links. The action space is modelled offline by using different time splits. The adaptation of the considered learning optimal policy through online learning is introduced to deal with the change of the environment. The simulation results show the effectiveness of the proposed adaptive learning algorithm.