Search

In the majority of data mining tasks, feature selection serves as an essential pre-processing step. The most important attributes are selected to lower the dimensionality reduction of data set and enhance the precision of classification. Natural heuristic algorithms are extensively employed in the realm of encapsulated feature selection. Based on the wrapper feature selection method, seven natural heuristic algorithms are used to solve feature selection problems and perform performance comparison, which include Slime Mold Algorithm (SMA), Whale Optimization Algorithm (WOA), Harris Hawks Optimization Algorithm (HHO), Marine Predator Algorithm (MPA), Butterfly Optimization Algorithm (BOA), Cuckoo Search (CS) and Firefly Algorithm (FA). At the same time, performance tests are carried out on 21 standard UCI data sets to verily the functionality of various algorithms, and the convergence curves and accuracy boxplots of 7 natural heuristic algorithms on 21 data sets are given. The simulation outcomes were assessed utilizing the mean and standard deviation of fitness, as well as the number of chosen features, and the running time, with the optimal value in bold. By comparing the comprehensive performance indexes, MPA obtained the maximum mean fitness value in most data sets (16 data sets), followed by FA (6 data sets). SMA obtained the best performance and finds the minimum eigenvalues (20 data sets) in multiple data sets and has an advantage in computing time. [ABSTRACT FROM AUTHOR]

Deep learning models have achieved high performance across different domains, such as medical decision-making, autonomous vehicles, decision support systems, among many others. However, despite this success, the inner mechanisms of these models are opaque because their internal representations are too complex for a human to understand. This opacity makes it hard to understand the how or the why of the predictions of deep learning models. There has been a growing interest in model-agnostic methods that make deep learning models more transparent and explainable to humans. Some researchers recently argued that for a machine to achieve human-level explainability, this machine needs to provide human causally understandable explanations, also known as causability. A specific class of algorithms that have the potential to provide causability are counterfactuals. This paper presents an in-depth systematic review of the diverse existing literature on counterfactuals and causability for explainable artificial intelligence (AI). We performed a Latent Dirichlet topic modelling analysis (LDA) under a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework to find the most relevant literature articles. This analysis yielded a novel taxonomy that considers the grounding theories of the surveyed algorithms, together with their underlying properties and applications to real-world data. Our research suggests that current model-agnostic counterfactual algorithms for explainable AI are not grounded on a causal theoretical formalism and, consequently, cannot promote causability to a human decision-maker. Furthermore, our findings suggest that the explanations derived from popular algorithms in the literature provide spurious correlations rather than cause/effects relationships, leading to sub-optimal, erroneous, or even biased explanations. Thus, this paper also advances the literature with new directions and challenges on promoting causability in model-agnostic approaches for explainable AI. • A survey on model-agnostic counterfactual approaches for XAI. • A novel taxonomy for model-agnostic counterfactual approaches for XAI. • A set of properties for causability systems for XAI. • Opportunities and challenges for causability systems based on modelagnostic counterfactual approaches. [ABSTRACT FROM AUTHOR]

High-frequency bipolar high-voltage pulsed electric field can be used in tumor ablation to uniform the electric field distribution in the ablated region and inhibit muscle contraction effectively. Developing a high-frequency bipolar pulsed power generator (HBPPG) with wide range of specifications, such as voltage magnitude, waveform, repetition rate, and pulsewidth, is of great significance for promoting the clinical application of irreversible electroporation. Hence, in this article, a novel modular high-voltage HBPPG circuit topology is proposed for such application. Each module consists of an H-bridge and an isolated inductor. Theoretical analysis, simulation, and experimental results show that this generator perfectly combines the advantages of solid-state Marx and bridge circuit, with high stability and redundancy. The pulse frequency, pulse duration, and waveform can be flexibly adjusted by a controller software algorithm according to load parameters. The viability of the proposed HBPPG is demonstrated by PSpice simulation and equal-scale prototype. The key parameters of the developed HBPPG are as follows: voltage amplitude up to ±10 kV with 500 ns–5 μs pulsewidths, 500 kHz within the burst, and 10 kHz within the continuation limited by the input high-voltage dc power supply. All the pulse parameters can be programmed arbitrarily. [ABSTRACT FROM AUTHOR]

In this paper, a sequence decision framework based on the Bayesian search is proposed to solve the problem of using an autonomous system to search for the missing target in an unknown environment. In the task, search cost and search efficiency are two competing requirements because they are closely related to the search task. Especially in the actual search task, the sensor assembled by the searcher is not perfect, so an effective search strategy is needed to guide the search agent to perform the task. Meanwhile, the decision-making method is crucial for the search agent. If the search agent fully trusts the feedback information of the sensor, the search task will end when the target is "detected" for the first time, which means it must take the risk of founding a wrong target. Conversely, if the search agent does not trust the feedback information of the sensor, it will most likely miss the real target, which will waste a lot of search resources and time. Based on the existing work, this paper proposes two search strategies and an improved algorithm. Compared with other search methods, the proposed strategies greatly improve the efficiency of unmanned search. Finally, the numerical simulations are provided to demonstrate the effectiveness of the search strategies. [ABSTRACT FROM AUTHOR]

• The models, imaging algorithms, and applications are systematically compared, analyzed, and concluded in this paper to explore the differences between the PAD-DBF and FDD-DBF techniques and provide guidance on the application of the technology. • A large number of simulations are investigated to explore the imaging characteristics of two rotor aerodynamic dipole source imaging technologies under different testing conditions, including different analysis frequencies, axis directions, SNRs, and measured distances. • The two imaging techniques are applied to the semi-anechoic chamber test for rotating dipole sources and the actual BO105 rotor wind tunnel test, respectively, which is further used to compare and validate the imaging effectiveness and applicability of the two techniques. The acoustic imaging technology for rotating dipole sources has been one of the topics concerned. In recent years, two state-of-the-art imaging technologies for rotating dipole sources were proposed, including beamforming technologies based on the predefined axis direction (PAD-DBF) and frequency-domain deconvolution (FDD-DBF). However, a systematic comparison of these two imaging technologies does not currently exist. In this paper, the models and imaging algorithms are firstly systematically compared and analyzed to explore the differences between the two technologies and provide guidance on the application of the technology. Secondly, extensive simulations are investigated to analyze the performance of the two imaging technologies under different working conditions. Finally, the two imaging technologies are applied to the semi-anechoic chamber test for rotating dipole sources and the actual BO105 rotor wind tunnel test, respectively, which is further used to compare and validate the imaging effectiveness and applicability of the two technologies. The analysis shows that both technologies can effectively achieve rotating dipole source localization. The PAD-DBF technology can achieve a more accurate rotating dipole source location, higher imaging resolution at a lower time cost, and it can get more robust imaging results in the low signal-to-noise ratio (SNR) of the wind tunnel environment. In contrast, the FDD-DBF technology can get a more accurate sound source magnitude while effectively imaging the rotating dipole source. [ABSTRACT FROM AUTHOR]

The exchanged crossed cube, denoted by E C Q (s , t) , is a novel interconnection network with fewer edges and smaller diameter compared to other variations of the corresponding hypercube. The linear array, denoted by L n , is one of the most popular topologies in optical networks. This paper addresses the routing and wavelength assignment for realizing E C Q (s , t) communication pattern on wavelength division multiplexing (WDM) optical network L n , where n = s + t + 1. We prove that the congestion for E C Q (s , t) on L n is equal to 2 s + t − 1 + ⌊ 2 t / 3 ⌋ , which is the lower bound of the minimum number of required wavelengths. In addition, an embedding scheme and an optimal wavelength assignment algorithm that achieve this bound are also proposed. [ABSTRACT FROM AUTHOR]

In this paper, we investigate the problem of minimizing energy cost for distributed Internet data centers (IDCs) in smart microgrids while taking system dynamics into consideration. Specifically, IDC operators expect to minimize the long-term energy cost with the uncertainties in electricity price, workload, renewable energy generation, and power outage state. At first, we formulate the problem as a stochastic program that captures service request distribution, server provisioning, energy storage management, generator scheduling, power transactions between smart microgrids, and main grids. Second, we use the Lyapunov optimization technique to design an operation algorithm, which enables an explicit tradeoff between energy cost saving and battery investment cost. Finally, the effectiveness of the proposed algorithm is evaluated with practical data. [ABSTRACT FROM PUBLISHER]

[Display omitted] • Algorithm for unmixing the flow spectra with partly known components. • Gauss model to fit the spectrum of the unknown component. • GA to optimize the parameters of the Gauss models. • Verify the algorithms by the simulated and experimental cyanobacteria data. Spectral unmixing algorithm is one of the key technologies for spectral flow cytometer in biology, chemistry and medicine. The proposed algorithm can separate the overlapping spectra automatically without the premeasured single stained or un-stained samples as the basic pure spectra. Genetic algorithm is adopted to search the optimal positions and peak sharps of the basic spectra derived from the unknown components, and then the concentration of each component can be estimated simultaneously by least squares method. Compared with conventional methods, the proposed algorithm has a wider application scope, such as the multi-stained samples with unknown components or the samples with auto-fluorescence. In the simulation, the convergence rate, accuracy and stability of the proposed algorithm are evaluated under the conditions of completely and partly unknown components. In the experiment, the flow spectra of cyanobacteria are processed, and the results demonstrate the feasibility and effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

• Acoustic inverse problems in the framework of ADMM are investigated. • The iteration algorithms of ADMM in the complex domain have been derived. • ADMM has been systematically compared with the sparse holography. • Both simulation and experimentation have verified the performance of ADMM. Acoustic imaging is an important technique for sound source identification which is widely used in noise control and fault detection. It can be described as a method for reconstructing the amplitude and phase distribution of an acoustic field in a given plane. When the size of radiating object is huge and measurement environment is noisy, accuracy and efficiency of acoustic imaging algorithm deteriorate easily. In this paper, a novel computational framework is investigated to simplify the algorithm and balance computation speed with accuracy. It converts acoustic imaging into an inverse problem and then divides this problem into two parts including forward model and denoising model. This computation framework is based on the alternating direction method of multipliers, which makes parallel implementation and distributed optimization possible, especially for large scale measurements. It is assumed that the acoustic field is composed of a cluster of monopole sources and the equivalent source method is taken to solve the acoustic inverse problem. Both L 2 -norm penalization based on the minimal energy assumption and L p -norm (0 < p ⩽ 1) penalization based on the sparsity hypothesis are taken in this framework. Thus, the regularization parameter in the algorithm, which balances the standard least square error and the norm of the source solution, can be coupled with some pre-existing methods incluing L-curve, GCV and Bayesian strategies. Different regularization parameters are proposed and the comparison is made in this paper. The proposed method provides a general and convenient method for different type of sound source priors and it is validated with an experiment. [ABSTRACT FROM AUTHOR]

The single-pixel imaging (SPI) technique is able to capture two-dimensional (2D) images without conventional array sensors by using a photodiode. As a novel scheme, Fourier single-pixel imaging (FSI) has been proven capable of reconstructing high-quality images. Due to the fact that the Fourier basis patterns (also known as grayscale sinusoidal patterns) cannot be well displayed on the digital micromirror device (DMD), a fast FSI system is proposed to solve this problem by binarizing Fourier pattern through a dithering algorithm. However, the traditional dithering algorithm leads to low quality as the extra noise is inevitably induced in the reconstructed images. In this paper, we report a better dithering algorithm to binarize Fourier pattern, which utilizes the Sierra–Lite kernel function by a serpentine scanning method. Numerical simulation and experiment demonstrate that the proposed algorithm is able to achieve higher quality under different sampling ratios. [ABSTRACT FROM AUTHOR]

Traffic data provide the basis for both research and applications in transportation control, management, and evaluation, but real-world traffic data collected from loop detectors or other sensors often contain corrupted or missing data points which need to be imputed for traffic analysis. For this end, here we propose a deep learning model named denoising stacked autoencoders for traffic data imputation. We tested and evaluated the model performance with consideration of both temporal and spatial factors. Through these experiments and evaluation results, we developed an algorithm for efficient realization of deep learning for traffic data imputation by training the model hierarchically using the full set of data from all vehicle detector stations. Using data provided by Caltrans PeMS, we have shown that the mean absolute error of the proposed realization is under 10 veh/5-min, a better performance compared with other popular models: the history model, ARIMA model and BP neural network model. We further investigated why the deep leaning model works well for traffic data imputation by visualizing the features extracted by the first hidden layer. Clearly, this work has demonstrated the effectiveness as well as efficiency of deep learning in the field of traffic data imputation and analysis. [ABSTRACT FROM AUTHOR]

Backlight power-saving algorithms can reduce the power consumption of the display by adjusting the frame pixels with optimal clipping points under some tradeoff criteria. However, the computation for the selected clipping points can be complex. In this paper, a novel algorithm is created to reduce the computation time of the state-of-the-art backlight power-saving algorithms. If the current frame is similar to the previous frame, it is unnecessary to execute the backlight power-saving algorithm for the optimal clipping points, and the derived clipping point from the previous frame can be used for the current frame automatically. In this paper, the motion vector information was used as the measurement of the similarity between adjacent frames, where the generation of the motion vector information requires no extra complexity since it is generated to reconstruct the decoded frame pixels before the display. The experiments showed that the proposed work can reduce the running time of the state-of-the-art methods by 25.21% to 64.22%, while the performances are maintained; the differences with the state-of-the-art methods in PSNR are only 0.02~1.91 dB, and those in power are only −0.001~0.008 W. [ABSTRACT FROM AUTHOR]

Raman spectroscopy is a non-destructive testing technology that has been widely used in biomedicine and other fields. However, the spontaneous Raman signal is weak. Therefore, the data collection is very time consuming, which has seriously hindered the application of Raman imaging technology in the fast dynamic system. At present, the reconstruction of Raman spectra based on multi-channel measurements is an effective method to solve this problem. However, the limitation of this algorithm is that the substance to be tested is known. In this paper, we have completed a new algorithm to solve this problem. Firstly, the feature abstraction of data is obtained based on the principal component analysis (PCA). Combining with the support vector machine (SVM), a spectral recognition algorithm is established. Secondly, the partial least squares and BP neural network are used to complete three algorithms for the reconstruction, which is based on the finite-dimensional model. Finally, the full Raman spectra and image are rapidly reconstructed. In the experiment, the Polymethyl Methacrylate (PMMA) is selected as the experimental sample. Meanwhile, the normalization algorithm and the polynomial regression are also introduced to complete the reconstruction of Raman spectrum. Based on the root mean square error (RMSE) and imaging time, the performance of the proposed algorithm is evaluated. The results show that the reconstructed spectrum by our algorithm in this paper is better than the traditional algorithms. Meanwhile, the imaging time is much lower than the traditional imaging modes. Therefore, this algorithm provides a theoretical support for the application of Raman imaging technology in fast dynamic systems. • Overcome the limitation of traditional reconstruction algorithm. • Reduce the influence of nonlinear factors. • Reduce the influence of the bad samples. • Improve the accuracy of reconstructed spectra. [ABSTRACT FROM AUTHOR]

A novel prediction approach based on the hidden Markov model and improved particle filtering for numerical control rotary table was proposed to estimate the degradation trend of repetitive positioning accuracy. Here, the resampling algorithm of the particle filter is improved and data-driven methods are introduced for the first time to study accuracy. The vibration signal obtained from an accelerated accuracy degradation test was selected as the test data. First, the original signal was de-noised and reconstructed by ensemble empirical mode decomposition-principal component analysis. Second, a hidden Markov model was trained by an observation matrix, which was composed of statistical characteristic values. Then, an early diagnosis of repetitive positioning accuracy degradation was obtained and the health status indicator of accuracy was built. Finally, the degradation trend model of repetitive positioning accuracy was established by improved particle filtering, and the residual accuracy life can also be calculated. With model calculations and experimental measurements, the results show that the approach is effective for numerical control rotary tables to estimate the degradation trend of repetitive positioning accuracy and residual accuracy life. [ABSTRACT FROM AUTHOR]

• Districting consideration is successfully integrated with the multi-customer JRP. • This study is vital to a company outsourcing logistics to a 3PL service provider. • Our numerical experiments are based on a real-life example in a bank system. • An effective GA-based algorithm is proposed for solving the problem. • Managerial insights provide baseline information before contract negotiation. This paper studies a multi-customer joint replenishment problem with districting consideration (MJRPDC) which is of particular importance to a company that outsources its transportation and delivery operations to a third-party logistics (3PL) service provider. To solve the problem, we first propose an innovative search algorithm for solving the traditional multi-customer joint replenishment problem in a given zone. Then we design a GA-based framework to handle the corresponding districting problem based on the performance of each district evaluated by using the proposed search algorithm. The proposed methodologies are demonstrated by using an example of solving MJRPDC for a bank. [ABSTRACT FROM AUTHOR]

The mechanical properties of the geo-materials are greatly affected by the internal micro structure system. Multiscale modelling and analysis are considered as an effective tool for revealing the mechanical process. Aiming at the problems of the light intensity decrease and the uneven illumination, a Retinex scale optimized image enhancement algorithm is proposed, based on the light reflection model and Retinex theory. Based on the Retinex theory, a Monte-Carlo random geometric numerical model of the microstructure is established, with some soil specimens in South China as the study object. After that, three-axis consolidation process of soft soil materials are simulated, which are compared with the experimental data. The results showthat: the Retinex scale optimized image enhancement algorithm can accurately estimate the illumination component, eliminate the influence of the uneven illumination, improve the contrast of image and retain the details of the microstructure simultaneously. The random field model with this algorithm is simple and effective, the image becomes clearer, and the contrast ratio is improved, after using Retinex algorithm to enhance the CT image of rock and soil. The threshold segmentation of the enhanced image, and the fidelity between the enhanced image and the original image, are higher than that of the homomorphic filtering method and histogram equalization method, which explained that the algorithm could preserve the microstructure information of the micro image well. The result of numerical simulation is similar with the one obtained from conventional three axis consolidation test, which proves that the simulation result is reliable. [ABSTRACT FROM AUTHOR]

Raman spectroscopy has shown great potential in biomedical applications. However, slow data acquisition of Raman spectra has seriously hindered the expansion of its application. In this paper, we have completed the reconstruction of Raman spectra with multi-channel measurements based global weighted linear regression. This algorithm establishes a linear regression function by optimizing the training samples and making a global assignment weighted to the optimizing samples. Simultaneously, the normalization and polynomial regression are introduced in order to improve the accuracy of reconstructed spectra. It has evaluated the Raman spectra of several materials. According to the root mean square error, the fitness between reconstructed and original spectra is excellent. This algorithm can be used in quickly testing for potential sample component in a substance, where the sample component to be tested is known and provides a theoretical support for the application of Raman imaging technology in fast dynamic systems. • We propose a method for optimizing the training samples based on Tanimoto coefficient to delete the bad sample. • We perform polynomial regression on the multi-channel measurements in order to reduce the influence of nonlinear factors and introduce the normalization, cross-validation algorithms to improve the accuracy of reconstructed spectra. • We establish a linear regression function by optimizing the training samples and making a global assignment weight to the optimizing samples. • We verify the effectiveness and the feasibility of the methods by processing the Raman spectra of several materials. [ABSTRACT FROM AUTHOR]