Adaptive Sampling Algorithm Research Articles

Evaluating an adaptive sampling algorithm to assist soil survey in New South Wales, Australia

Knowledge of the spatial variation of soil is important in modern agricultural management. To attain this knowledge, ground-based samples are required in combination with many ground-based, air-borne and space-borne sensors from the Internet of Things. Compared to traditional grid and simple random sampling that are designed for fixed sensors, adaptive sampling is not well studied. In this study, we propose a prior-based adaptive sampling scheme to collect soil samples for estimation of ground-based Gamma-ray potassium across an 80-ha field in a semi-arid landscape, in New South Wales, Australia. We compare the performance of the sampling algorithm via a linear mixed model between various adaptive sampling schemes with prior information of varying quality (e.g. ground apparent electrical conductivity, air-borne Gamma-ray potassium, and a legacy map of clay content). We also compare the model performance of the adaptive sampling scheme with more conventional grid and simple random sampling schemes. Results show that the adaptive sampling scheme was superior to the grid and simple random sampling schemes in terms of the accuracy of the linear mixed model when the sampling size was small (<15 additional samples) due to the use of prior information. The accuracy of the linear mixed models associated with the adaptive sampling schemes deteriorated when the quality (correlation with the target soil variable) of the prior information decreases. We conclude that the algorithm has the potential to be applied generally for automated adaptive sampling design (e.g., on an autonomous vehicle) when sampling cost is large and travelling time of the sensor is relatively small.

Ranking of Ligand Binding Kinetics Using a Weighted Ensemble Approach and Comparison with a Multiscale Milestoning Approach.

To improve lead optimization efforts in finding the right ligand, pharmaceutical industries need to know the ligand's binding kinetics, such as binding and unbinding rate constants, which often correlate with the ligand's efficacy in vivo. To predict binding kinetics efficiently, enhanced sampling methods, such as milestoning and the weighted ensemble (WE) method, have been used in molecular dynamics (MD) simulations of these systems. However, a comparison of these enhanced sampling methods in ranking ligands has not been done. Hence, a WE approach called the concurrent adaptive sampling (CAS) algorithm that uses MD simulations was used to rank seven ligands for β-cyclodextrin, a system in which a multiscale milestoning approach called simulation enabled estimation of kinetic rates (SEEKR) was also used, which uses both MD and Brownian dynamics simulations. Overall, the CAS algorithm can successfully rank ligands using the unbinding rate constant koff values and binding free energy ΔG values, as SEEKR did, with reduced computational cost that is about the same as SEEKR. We compare the CAS algorithm simulations with different parameters and discuss the impact of parameters in ranking ligands and obtaining rate constant and binding free energy estimates. We also discuss similarities and differences and advantages and disadvantages of SEEKR and the CAS algorithm for future use.

A predictive model of recreational water quality based on adaptive synthetic sampling algorithms and machine learning

Predicting recreational water quality is one of the most difficult tasks in water management with major implications for humans and society. Many data-driven models have been used to predict water quality indicators to allow a real time assessment of public health risk. This assessment is most commonly based on Faecal Indicator Bacteria (FIB), with the value of FIB compared with thresholds published in guidelines. However, FIB values usually tend to be unbalanced within water quality datasets, with small proportions of data exceeding guideline thresholds and far larger numbers that do not. This can be a limiting factor in the uptake of model predictions since, even if the overall accuracy is high, the sensitivity of the predictions can be low. To address this issue, this paper proposes an adaptive synthetic sampling algorithm (ADASYN) to generate synthetic above-threshold FIB instances and test the validity of the approach for the prediction of recreational water quality. The models in this paper are based on four machine learning techniques: k-mean nearest neighbour, boosting decision tree, support vector machine, and multi-layer perceptron artificial neural network and are applied to five different locations in Auckland, New Zealand. Aside from support vector machine, all models provide favourable predictions with relatively high sensitivity (around 75%) and overall accuracy (over 90%), indicating that both the compliant and exceedance conditions can be effectively predicted through the use of more sophisticated model training which involves artificial data. Considering the model accuracy and stability, boosting decision trees (BDT) and multi-layer perceptron artificial neural (MLP-ANN) network are the best two models and the multi-layer perceptron is the most efficient with the shortest computation time.

Bayesian identification and model comparison for random property fields derived from material microstructure

This paper provides a Bayesian framework for determining the spatial variability of the apparent material properties of two-phase composites. Bayesian analysis is applied to learn the parameters of the random fields that characterize the mesoscale material properties using computer-simulated images of the material microstructure. The information from such images is used to define the likelihood function of the random field parameters given the homogenized microscale data. The uncertainty in the parameter estimates is quantified through determining their full posterior distribution instead of a point estimate by application of an adaptive sampling algorithm. In addition, the Bayesian approach allows for assessing the plausibility of different correlation models belonging to the Matérn class through computing their respective marginal likelihoods for various choices of the smoothness parameter. The framework is applied to a composite material with large inclusion/matrix stiffness ratio and useful conclusions are derived with regard to the most appropriate correlation model and its parameters.

Modeling the 3-D structure of ignition experiments at the NIF

This work details a model used to infer the 3-D structure of the stagnated hot-spot and shell of inertial confinement fusion implosion experiments at the National Ignition Facility. The model assumes that 3-D low-mode drive perturbations can account for the majority of stagnation asymmetries experimentally observed. It uses an adaptive sampling algorithm to navigate the 24-D input parameter space to find a 3-D x-ray flux asymmetry whose application to an otherwise symmetric implosion results in a consistent match between synthetic and experimental diagnostic observables. The model is applied to a series of experiments and is able to achieve a consistent match for over 41 different observables, providing a high-fidelity reconstruction of the stagnation hot-spot and shell profile.

Ranking of Ligand Binding Kinetics using a Weighted Ensemble Approach and Comparison with Milestoning

In order to improve lead optimization efforts in finding the right ligand, pharmaceutical industries need to know the ligand's binding kinetics, such as binding and unbinding rate constants, which are known to be good indicators of the ligand's efficacy in vivo. To predict binding kinetics quickly, enhanced sampling methods, such as milestoning and the weighted ensemble (WE) method, have been used in molecular dynamics (MD) simulations of these systems. However, a direct comparison of the enhanced sampling methods in ranking ligands has not been done. Hence, a WE approach called the concurrent adaptive sampling (CAS) algorithm was used to rank seven ligands for beta-cyclodextrin, a system in which a milestoning approach called simulation enabled estimation of kinetic rates (SEEKR) was also used. Overall, the CAS algorithm can successfully rank ligands using the unbinding rate constants koff's and binding free energies, as SEEKR did, with reduced computational cost that is about the same as SEEKR. We discuss "best practices" for using WE approaches like the CAS algorithm and pros and cons of each enhanced sampling method for future use.

An Efficient Adaptive Importance Sampling Method for SRAM and Analog Yield Analysis

Performance failure has become a major threat for various memory and analog circuits. It is challenging to estimate the extremely small failure probability when failed samples are distributed in multiple disjoint regions. In this article, we propose an adaptive importance sampling (AIS) algorithm. AIS has several iterations of sampling region adjustments, while existing methods predecide a static sampling distribution. We design two adaptive frameworks based on resampling and population Metropolis–Hastings (MH) to iteratively search for failure regions. The experimental results of the AIS method exhibit better efficiency and higher accuracy. For SRAM bit cell with single failure region, the AIS method uses 2– $27{\times }$ fewer samples and reaches better accuracy when compared to several recent methods. For a two-stage amplifier circuit with multiple failure regions, the AIS method is $90{\times }$ faster than Monte Carlo and 7–23 ${\times }$ over other methods. For charge pump circuit and $C^{2}MOS$ master–slave latch circuit, the AIS method can reach 6– $18{\times }$ and 4– $6{\times }$ speedup over other methods, respectively.

Rainfall Monitoring Based on Machine Learning by Earth-Space Link in the Ku Band

Recently, the oblique earth-space links (OELs) between satellite and earth station have been used for rainfall monitoring as a supplement to existing observation methods. Most recent studies achieved the rainfall measurement by OELs based on the empirical method such as power-law (PL) model. In practice, two crucial issues need to be addressed: 1) identification of rain and no-rain periods; and 2) determination of attenuation baseline. To solve these problems, this article adopts several machine learning algorithms based on the analysis of earth-space link signal characteristics. For the first issue, we choose the support vector machine as a classifier and the adaptive synthetic sampling algorithm is deployed to eliminate the effects caused by the data imbalance. For the second issue, the long short-term neural network is selected for the determination of attenuation baseline since it has a good ability to solve time-series problem. In terms of the rainfall inversion, we establish a new model by combining the back-propagation (BP) network and genetic algorithm (GA). The PL model is also used as a comparison. To validate the proposed method, we set up an earth-space link that receives the signal from AsiaSat5 in 12.32 GHz. The results demonstrate that the two issues are successfully addressed and the inversion of precipitation is also carried out. Compared to disdrometer, the correlation and mean absolute error of GA-BP model are 0.83 and 1.30 mm/h, respectively, indicating a great potential to use densely OELs for global precipitation monitoring.

Classifying cardiac rhythms by means of digital signal processing and machine learning

Electrocardiogram (ECG) measures the electrical activity of the heart, which can be used in the diagnosis of different heart diseases. In the scientific literature there are many studies that have been applied machine learning for recognizing ECG patterns, where most of them attempt to classify heart beats. This paper presents a novel methodology for automatically classifying seventeen cardiac rhythms by means of digital signal processing and machine learning. The steps before the classification include the mapping of ECG signal to the frequency domain through power spectrum density, class balance with Adaptive Synthetic Sampling algorithm, and statistical normalization. The classifiers employed were Support Vector Machine, Multilayer Perceptron Neural Network, k-Nearest Neighbors, and Random Forest. The results showed accuracy, sensitivity, specificity, and Fleiss' kappa of up to 98.86%, 99.93%, 98.85%, and 89.68%, respectively, which are relatively better than the performance observed in the state-of-the-art works. In addition, this study highlighted that when the class balance procedure is applied, the classification step becomes less complex and can increase in terms of performance.

Learning-Based Beamforming for Multi-User Vehicular Communications: A Combinatorial Multi-Armed Bandit Approach

There is an urgent need to support high data rate among connected vehicles for both safety and infotainment purposes. However, high rate for vehicular communications is impacted by several challenges such as optimal beam tracking for multiple simultaneous-transmitting vehicles in highly-dynamic environments. In this paper, we aim to maximize the overall network throughput for multi-vehicular communications. We propose a reinforcement learning (RL) approach called combinatorial multi-armed bandit (CMAB) framework, where multiple arms (i.e., actions) form a super arm and can be played together, to handle the beam selection problem in a vehicular network. More specifically, we propose an adaptive combinatorial Thompson sampling algorithm, namely adaptive CTS, that CMAB embodies for appropriate selection of simultaneous beams in a high-mobility vehicular environment. The proposed approach applies the smart exploration-exploitation trade-off for the fast selection of beams. As the proposed adaptive CTS scheme produces a higher complexity over the search space which increases exponentially with the number of user, we also propose a sequential Thompson sampling (TS) algorithm where beam selection is performed in a one-by-one manner. We analyze the regret bounds for the proposed beam tracking algorithms. The performance of the proposed strategies are evaluated for multi-vehicle millimeter-wave (mmWave) simulation environments. Our results suggest that the proposed sequential approach performs almost similar to the simultaneous adaptive CTS scheme for tracking optimal beams in a multi-vehicular network with much reduced complexity. Simulation results also show that both of our proposed strategies approach the optimal achievable rate achieved by the genie-aided solution.

PAADA: Physical-World-Aware Approximate Data Acquisition for Image Sensor Networks

To observe the complicated physical world, sensor networks are widely used for data collection. Moreover, due to the limited energy, storage and computation capacity of sensor nodes, approximate data acquisition in adaptive sampling manner is a wide choice. Nevertheless, existing data acquisition methods are most designed for univariate data (e.g., temperature), and thus not applicable to image data with high dimensions and complex structures. In this paper, we propose a framework of Physical-world-Aware Adaptive Data Acquisition (PAADA) for image sensor networks, to sample data adaptively with pre-specified error bound. First, based on the convolutional autoencoders (CAEs), PAADA compresses the high-dimensional image data into a feature vector with a handful of hidden variables which compactly capture the key features of the image data. Second, PAADA designs a Physical-world-aware Adaptive Sampling (PAS) algorithm based on the Hermitee interpolation. Under the feature space, the PAS algorithm adjusts the sampling frequency automatically by considering the change trend of the feature vector. In addition, the feature vectors at non-sampling time points can be recovered with O (∈) approximation guarantee to the ground truths. Next, PAADA recovers the image data at non-sampling time points based on the recovered feature vectors. Finally, for each sensor, PAADA returns an image series composed of sampled images (at sampling time points) and approximate images (at non-sampling time points). Experiments on real-world datasets demonstrate that the proposed PAADA has high performance in both accuracy and energy consumption.

A Novel Wireless Network Intrusion Detection Method Based on Adaptive Synthetic Sampling and an Improved Convolutional Neural Network

The diversity of network attacks poses severe challenges to intrusion detection systems (IDSs). Traditional attack recognition methods usually adopt mining data associations to identify anomalies, which has the disadvantages of a high false alarm rate (FAR), low recognition accuracy (ACC) and poor generalization ability. To ameliorate the comprehensive capabilities of IDS and strengthen network security, we propose a novel intrusion detection method based on the adaptive synthetic sampling (ADASYN) algorithm and an improved convolutional neural network (CNN). First, we use the ADASYN method to balance the sample distribution, which can effectively prevent the model from being sensitive to large samples and ignore small samples. Second, the improved CNN is based on the split convolution module (SPC-CNN), which can increase the diversity of features and eliminate the impact of interchannel information redundancy on model training. Then, an AS-CNN model mixed with ADASYN and SPC-CNN is used for intrusion detection tasks. Finally, the standard NSL-KDD dataset is selected to test AS-CNN. The simulation illustrates that the accuracy is 4.60% and 2.79% higher than that of the traditional CNN and RNN models, and the detection rate (DR) increased by 11.34% and 10.27%, respectively. Additionally, the FAR decreased by 15.58% and 14.57%, respectively, compared with the two models.

Oil Plume Mapping: Adaptive Tracking and Adaptive Sampling From an Autonomous Underwater Vehicle

We have developed an adaptive sampling algorithm for an Explorer autonomous underwater vehicle (AUV) to conduct in-situ analysis of acoustic measurements to perform autonomous oil plume detection and tracking. The methodology of the tracking phase involves ongoing analysis of the detected plume, assessing target validity and proximity for AUV decision-making for plume mapping. We previously introduced the bumblebee flight path, a new biomimetic search pattern designed to maximize the spatial coverage in the oil plume detection phase. This paper focuses on a new tracking strategy as the key adaptive stage in our plume delineation. For initial development we used a 360-degree scanning sonar sensor model. Simulations were done with different plume models to assess the performance of the developed adaptive sampling algorithm. A convergence study demonstrated that the algorithm could successfully track the boundary of a non-regular shaped/patchy oil plume at up to a 0.05Hz sampling frequency. A sensitivity study identified the correlations between plume feature complexity and the anticipated range of acoustic measurement update delays. The decision-making architecture consists of three separate components which implements either proximity or boundary following control and contributes to the final decision on the next desired heading of the vehicle. A weight ratio, that determined the relative allocation of each component, was varied to study its impact on the tracking performance of the AUV. The novelty of our approach is in addressing the discontinuous and patchy nature of realistic oil plumes. Our sampling algorithm and its performance in simulations is a significant step beyond the practical limitations of existing gradient-following methods because it accounts for the oil patches and droplets which gradient-following algorithms do not.

User-Driven Adaptive Sampling for Massive Internet of Things

Energy conservation techniques are crucial to achieving high reliability in the Internet of Things (IoT) services, especially in the Massive IoT (MIoT), which stringently requires cost-effective and low-energy consumption for battery-powered devices. Most of the proposed techniques generally assume that data acquiring and processing consume significantly lower than that of communication. Unfortunately, this assumption is incorrect in the MIoT scenario, which mostly involves the low-power wide-area network (LPWAN) and complex data sensing operations (e.g., biological and seismic sensing) using “power-hungry” sensors (e.g., gas sensors, seismometers). Thus, sensing actions may consume even more energy than transmission. In addition, none of them support end-users in controlling the trade-off between energy conservation and data precision. To deal with these issues, we propose an adaptive sampling algorithm that estimates the optimal sampling frequencies in real-time for IoT devices based on the changes of collected data. Given a user's saving desire, our algorithm could minimize the device's energy consumption while ensuring the precision of collected information. Practical experiments over IoT datasets have shown that our algorithm can reduce the number of acquired samples up to 20 times compared with a traditional fixed-rate approach at extremely low Normal Mean Error value around 3.45%.

Surrogate Model of Catalytic Reforming Process Based on a New Adaptive Sampling Algorithm

Surrogate Model of Catalytic Reforming Process Based on a New Adaptive Sampling Algorithm

Simulation-Based Algorithms for Markov Decision Processes: Monte Carlo Tree Search from AlphaGo to AlphaZero

AlphaGo and its successors AlphaGo Zero and AlphaZero made international headlines with their incredible successes in game playing, which have been touted as further evidence of the immense potential of artificial intelligence, and in particular, machine learning. AlphaGo defeated the reigning human world champion Go player Lee Sedol 4 games to 1, in March 2016 in Seoul, Korea, an achievement that surpassed previous computer game-playing program milestones by IBM’s Deep Blue in chess and by IBM’s Watson in the U.S. TV game show Jeopardy. AlphaGo then followed this up by defeating the world’s number one Go player Ke Jie 3-0 at the Future of Go Summit in Wuzhen, China in May 2017. Then, in December 2017, AlphaZero stunned the chess world by dominating the top computer chess program Stockfish (which has a far higher rating than any human) in a 100-game match by winning 28 games and losing none (72 draws) after training from scratch for just four hours! The deep neural networks of AlphaGo, AlphaZero, and all their incarnations are trained using a technique called Monte Carlo tree search (MCTS), whose roots can be traced back to an adaptive multistage sampling (AMS) simulation-based algorithm for Markov decision processes (MDPs) published in Operations Research back in 2005 [Chang, HS, MC Fu, J Hu and SI Marcus (2005). An adaptive sampling algorithm for solving Markov decision processes. Operations Research, 53, 126–139.] (and introduced even earlier in 2002). After reviewing the history and background of AlphaGo through AlphaZero, the origins of MCTS are traced back to simulation-based algorithms for MDPs, and its role in training the neural networks that essentially carry out the value/policy function approximation used in approximate dynamic programming, reinforcement learning, and neuro-dynamic programming is discussed, including some recently proposed enhancements building on statistical ranking &amp; selection research in the operations research simulation community.

Application of an adaptive MCMC method for the heat flux estimation

ABSTRACTBased on Bayesian inference, an adaptive single component Metropolis–Hastings (SCMH) sampling algorithm is presented to estimate the surface heat flux which is a typical inverse heat conduction problem. The surface heat flux is expressed by two different basis functions, the piecewise linear function and the Fourier series, and for speeding up convergence process, the parameters of the Fourier series are arbitrarily decoupled in the burn-in period. For validating and analysing the performance of the sampling algorithm, several cases are discussed. The results show that the adaptive SCMH sampling algorithm is feasible and has accuracy comparable to the conjugate gradient method (CGM), and decoupling method can successively shorten the burn-in period. It also can be found that ‘adaptive’ algorithm can obviously accelerate the convergence process. Furthermore, the performances of the two basis functions are also analysed. Because of the correlation, the paths of Markov Chain of the Fourier coefficients seem more ‘random’, which indicated that the algorithm with Fourier series is more efficient than that with piecewise linear function. The influence of the number of parameters is also studied, which is similar as the regularization of the CGM, the fewer the parameters is, the smoother the estimated heat flux becomes.

Foldamer Architectures of Triazine-Based Sequence-Defined Polymers Investigated with Molecular Dynamics Simulations and Enhanced Sampling Methods.

Triazine-based sequence-defined polymers have recently been developed that are biomimetic and robust. In molecular dynamics (MD) simulations, the triazine polymers were shown to form linear nanorod foldamers through hydrogen bonding and π-π interactions. The nanorod foldamers have motifs resembling those of DNA, α-helices, and β-sheets and have potential to be useful building blocks for new macromolecules and materials. To understand the formation of nanorod foldamers, we investigate how linker structures in the middle of the triazine polymers lead to folding using MD simulations. We found that a variety of linkers can participate in folding but that specific linker structures are more favorable than others, depending on the polymer length. Folding of hexamers into well-defined nanorod foldamers was most favorable with pentanediamine and ortho-xylenediamine linkers in the center of the polymers. Foldamers with ortho-xylenediamine linkers in the center were investigated for longer polymers, i.e., octamers and decamers, using two different enhanced sampling methods, since regular MD simulations had failed to show any folding for these longer polymers. In particular, the recently developed concurrent adaptive sampling (CAS) algorithm and replica exchange molecular dynamics (REMD) were used. We found that the two enhanced sampling methods did lead to the observation of foldamers and that REMD revealed new foldamer architectures where cis-trans isomerizations had occurred. Foldamer formation, diversity, and the strengths and limitations of simulation techniques are discussed. These findings provide new insights into the diversity of foldamer architectures for a new type of biomimetic synthetic polymer.

Energy modeling and adaptive sampling algorithms for energy‐harvesting powered nodes with sampling rate limitations

AbstractThis article explores the implementation of different sampling strategies for a practical energy‐harvesting wireless device (sensor node) powered by a rechargeable battery. We look for a realistic yet effective sampling strategy that prevents packet delivery failures, which is simple enough to be implemented in low‐complexity hardware. The article proposes methods that balance erratic energy arrivals and include advantages of dynamic data‐driven approaches based on historical data. Due to the industrial requirements in terms of minimum acceptable sampling frequency, we also integrate sampling rate limits and verify the proposed methods. To do so, we simulated the operation of an industrial data logger powered with a solar panel relying on the enhanced state of the model for battery charging. Finally, the proposed methods are compared based on energy consumption over a year and the amount of packet delivery failures, thus showing how some modifications of available strategies achieve satisfactory performance in this sense.

Dynamic sampling rate algorithm (DSRA) implemented in self-adaptive software architecture: a way to reduce the energy consumption of wireless sensors through event-based sampling

With the recent digitalization trends in the industry, wireless sensors are, in particular, gaining a growing interest. This is due to the possibility of being installed in inaccessible locations for wired sensors. Although great success has already been achieved in this area, energy limitation remains a major obstacle for further advances. As such, it is important to optimize the sampling with a sufficient rate to catch important information without excessive energy consumption, and one way to achieve sufficient sampling is using adaptive sampling for sensors. As software plays an important role in the techniques of adaptive sampling, a reference framework for software architecture is important in order to facilitate their design, modeling, and implementation. This study proposes a software architecture, named Rainbow, as the reference architecture, also, it develops an algorithm for adaptive sampling. The algorithm was implemented in the Rainbow architecture and tested using two datasets; the results show the proper operation of the architecture as well as the algorithm. In conclusion, the Rainbow software architecture has the potential to be used as a framework for adaptive sampling algorithms, and the developed algorithm allows adaptive sampling based on the changes in the signal.