Generalized Reduction Method Research Articles

Machine learning a surrogate model for transmission loss on a Pekeris waveguide

The connection between ocean environmental parameters and transmission loss (TL) can be modeled using sound propagation models. We can analyze and improve these models using information theoretic tools such as the Fisher information matrix (FIM) and methods for model reduction, where we identify parameters that can be removed from a model without sacrificing accuracy. Critical to these methods for model reduction is the ability to evaluate derivatives of the model predictions of TL with respect to model parameters. Calculating derivatives of TL models using finite difference methods with sufficient accuracy is challenging due to the complexity of sound fields in an ocean environment; instead, we train a surrogate machine learning model to which we can apply automatic differentiation (AD). We propose a general method for model reduction in which a ML model is used as a surrogate model for physical sound propagation ocean models, which is then used to find potential simplifications of the original model. We demonstrate this method using the Pekeris waveguide model of the ocean, which calculates underwater acoustic TL in an ocean environment due to seafloor characteristics. [Work supported by Office of Naval Research]

A Review of the Accuracy of Primal Assembly Model Order Reduction Techniques

Model order reduction (MOR) techniques have been used in structural dynamics for over sixty years as a quick and effective way to examine the dynamic behavior of complex finite element (FE) models. Various macroscopic phenomena arise from the physics and mechanics of the underlying microstructure, and MOR approaches can be employed in conjunction with the multiscale computational homogenization approach, which is deriving local macroscopic constitutive responses from underlying microstructure; to predict the dynamic behavior of heterogeneous macrostructures. The accuracy and efficiency of the most representative approaches from dynamic condensation, component mode synthesis (CMS), and general model reduction (GMR) methods are compared in this paper. The techniques considered are Guyan and the improved reduced system (IRS) from classical dynamic condensation, Craig-Bampton (CB) and enhanced Craig-Bampton (ECB) from CMS, IRS-based substructuring, GMR, and GMRPlus. All the substructuring methods employ primal assembly at the boundary interface of the substructures. The accuracy and computing efficiency of the techniques are compared in-depth using real-world engineering problems, and the effect of the number of substructures on the accuracy and computational cost is also examined.

Feature Extraction of High-dimensional Data Based on J-HOSVD for Cyber-Physical-Social Systems

With the further integration of Cyber-Physical-Social systems (CPSSs), there is explosive growth of the data in CPSSs. How to discover effective information or knowledge from CPSSs big data and provide support for subsequent learning tasks has become a core issue. Moreover, modern applications in CPSSs increasingly rely on the processing and analysis of high-dimensional data; the correlation and internal structure of these high-dimensional data are gradually becoming more complex, which further makes traditional machine learning algorithms a little inadequate in processing these data. In this article, we propose two general dimension reduction and feature extraction methods for high-dimensional data based on joint tensor decomposition, namely core feature extraction methods and factor feature extraction methods, which can effectively mine out the common components and hidden patterns of high-dimensional data by joint analysis while maintaining the original data structure. We also verified the effectiveness of the methods from both theoretical and practical aspects. Furthermore, we extend the two feature extraction methods to the tensor distance scenario and illustrate that the compressed features extracted by our models can keep the global information of original data well. Finally, we evaluated proposed methods on two benchmark datasets through classification tasks, and experimental results show that the low-dimensional features extracted by the proposed models have higher classification accuracy than the direct classification of the original data, which further verifies the effectiveness and robustness of our methods.

Ultralow detection limit MEMS hydrogen sensor based on SnO2 with oxygen vacancies

Controllable and efficient construction of oxygen vacancies on the surface of metal oxide semiconductors (MOSs) is essential for their application in the gas sensor. Herein, a general H2 reduction method is developed to synthesize SnO2 with oxygen vacancies defect (SnO2-D) by annealing the SnO2 in a H2 atmosphere at different temperature (300 °C, 400 °C and 500 °C), and then named SnO2-D3, SnO2-D4 and SnO2-D5, respectively. It was found that although the determined specific surface areas for pristine SnO2, SnO2-D3, SnO2-D4 and SnO2-D5 are 103.749 m2g−1, 63.316 m2g−1, 47.652 m2g−1 and 15.541 m2g−1, respectively, the gas sensitivity test results indicate that the SnO2-D4 Micro-Electro-Mechanical System (MEMS) sensor shows improved response and excellent low-concentration detection capability (down to 0.1 ppm) to H2 compared with that fabricated with pristine SnO2, SnO2-D3 and SnO2-D5. The abnormal relationship between specific surface area and gas sensing performance is attributed to the more oxygen vacancies of SnO2-D4 surface. In addition, ZnO (ZnO-D) and In2O3 (In2O3-D) with oxygen vacancy defects based on the H2 reduction method show better gas sensitivity than ZnO and In2O3 sensors, which further proves that oxygen vacancy defects can effectively improve the gas-sensing performance of MOSs.

Research on failure mechanism of pebble slope based on nonsynchronous double strength reduction method

This project is located on the gentle slope of the third-level terrace in the upper reaches of the Dadu River. Because of the deep ice water accumulation layer and the inclusion of a large number of pebbles and silt lenses, studying the failure mechanism of the slope of the deep heterogeneous pebble accumulation body is critical. The cohesive force and internal friction angle are nonsynchronously reduced through the software program by comparing the slope equivalent to the plastic strain area calculated by nonsynchronous reduction step conditions with the stress and strain data monitored on site. Simultaneously, the displacement and load data of the sample during the staged loading process, were calculated in detail. The results proved that during the predetermined load application process, the pebbles are gradually shifted, and the silt lens interspersed between the pebbles is quickly destroyed. The friction angle between the pebbles provides the shear strength, further verifying the rationality of the nonsynchronous dual-strength reduction method. Compared with the general synchronous reduction method, analyzing the slope stability is more reasonable. Simultaneously, the double strength reduction method provides a solution to calculate the slope stability, and can be used for similar slope projects in the future.

A General Order Reduction Method of Wideband Digital Predistortion Model Using Attention Mechanism

In wireless networks, for the common in‐phase and quadrature‐phase (I/Q) imbalance in the transmitters, the I/Q branch models of digital predistortion (DPD) need to be identified separately, to improve the linearization effects. The existing order reduction methods of the predistorter are based on the contributions of the complex basis function terms, so as not to deal with the different contributions of I/Q components of the complex basis function terms caused by the separate identification of the I/Q branch models. The separate pruning of the I/Q branch models will increase the complexity. Aiming at this issue, this paper proposes a general order reduction method based on the attention mechanism for the predistortion of the power amplifiers (PAs). This method is suitable for pruning both the traditional models and neural network‐based models. In this method, the attention mechanism is used to evaluate the contributions of the real basis function terms to the predistorted output’s I/Q components through offline training, and the influence of the cross terms of the I/Q branch models is considered. The experimental results based on the comparison with other typical methods under 100 MHz Doherty PA and different I/Q imbalance levels show that this method has superior pruning performance and good linearization ability.

Engineering sulfur vacancies in basal plane of MoS2 for enhanced hydrogen evolution reaction

Engineering sulfur vacancies in the basal plane of MoS2 is an effective method to enhance its catalytic activity. However, the traditional methods are only effective for multilayer or monolayer MoS2 supported on the substrate (such as Si/SiO wafer or carbon paper), which are not applied in bulk or commercial MoS2. In this work, we developed a simple and general chemical reduction method for engineering S-vacancies in the MoS2 basal plane and additional active edge sites, which can be exploited to improve the HER catalytic performance of MoS2. The optimized MoS2 nanosheets with S-vacancies exhibit excellent HER activity with a small overpotential of 190 mV at large density of 10 mA/cm2 and a low Tafel slope of 54 mV decade−1. Moreover, prominent electrochemical durability was also achieved. This finding proposes an effective strategy to fabricating S-vacancies for activating basal plane of 2D transition metal chalcogenides materials.

Stratification as a general variance reduction method for Markov chain Monte Carlo.

The Eigenvector Method for Umbrella Sampling (EMUS) [46] belongs to a popular class of methods in statistical mechanics which adapt the principle of stratified survey sampling to the computation of free energies. We develop a detailed theoretical analysis of EMUS. Based on this analysis, we show that EMUS is an efficient general method for computing averages over arbitrary target distributions. In particular, we show that EMUS can be dramatically more efficient than direct MCMC when the target distribution is multimodal or when the goal is to compute tail probabilities. To illustrate these theoretical results, we present a tutorial application of the method to a problem from Bayesian statistics.

Reducing noise from twin supersonic jets using very-low-frequency control

The intense jet noise radiated by closely spaced, twin supersonic hot jets leads to sound-induced structural vibration, fatigue and personnel related operational difficulties. Experimental, theoretical, and computational investigations into the physics and control of jet noise have identified several important sound sources, including wavepackets, screech, Mach wave radiation, and broadband shock associated noise (BBSAN). Reducing the loudest sources of jet noise, without sacrificing propulsive performance, has relied on intuition, parametric survey, or optimal control techniques. With the aim of developing a more general and robust method of jet noise reduction (JNR), we present a physics-based approach that leverages very-low-frequency jet dynamics in order to achieve JNR whilst preserving propulsive performance. Our approach formulates the control problem using the very-low-frequency global modes of the compressible Navier-Stokes operator linearized about the jet mean flow to disrupt the nonmodal transient growth processes. The presentation will showcase uncontrolled and controlled single and twin supersonic hot jets issuing from biconical nozzles, with conditions and geometries motivated by tactical Naval aircraft. The predictions utilize fully resolved numerical simulations whose data inform the control development and which evaluate its performance on the jet exhaust and on its radiated noise.

Synthesis of high-quality graphene with enhanced electrochemical properties by two-step reduction method

In this study, we reported a facile two-step reduction method to prepare a high-quality graphene with intact structure. Hydroxyl group is acquired by utilizing LiAlH4 to reduce the oxygen-containing functional groups in GO. Then it is further reduced by ethanethiol/aluminum trichloride (Et-SH/AlCl3) to obtain graphene. The results of Infrared and XPS showed that the prepared graphene material (rGO2) has been reduced completely and had structural integrity. Besides, the conductivity of rGO2 obtained can reach 2.632 * 103 S m−1, which is higher than that obtained by general chemical reduction method and high temperature reduction method. Investigated as a potential material for supercapacitors, the as-prepared graphene reveals excellent electrochemical properties, including good rate capability and cycling stability, which mainly benefits from its intact structural characteristics.

A general reduction method for fuzzy objective relation systems

Fuzzy objective relation systems are an important class of datasets that are generalizations of many types of decision tables. This paper proposes an approach, based on relation systems and fuzzy sets, to reduce data redundancy in fuzzy objective relation systems. We study lower and upper approximation reductions of a relation system for a given fuzzy set. As a generalization of such reductions, we consider lower and upper approximation reductions of fuzzy objective relation systems and give their corresponding reduction algorithms using methods based on the discernibility matrix. We note that the usual positive region reduction for a decision table can be considered a special case of our lower approximation reduction. Finally, we provide two examples from the UCI datasets to verify our theoretical results. These results can help in the decision making analysis of fuzzy objective relation systems.

Reduced Models of Point Vortex Systems.

Nonequilibrium statistical models of point vortex systems are constructed using an optimal closure method, and these models are employed to approximate the relaxation toward equilibrium of systems governed by the two-dimensional Euler equations, as well as the quasi-geostrophic equations for either single-layer or two-layer flows. Optimal closure refers to a general method of reduction for Hamiltonian systems, in which macroscopic states are required to belong to a parametric family of distributions on phase space. In the case of point vortex ensembles, the macroscopic variables describe the spatially coarse-grained vorticity. Dynamical closure in terms of those macrostates is obtained by optimizing over paths in the parameter space of the reduced model subject to the constraints imposed by conserved quantities. This optimization minimizes a cost functional that quantifies the rate of information loss due to model reduction, meaning that an optimal path represents a macroscopic evolution that is most compatible with the microscopic dynamics in an information-theoretic sense. A near-equilibrium linearization of this method is used to derive dissipative equations for the low-order spatial moments of ensembles of point vortices in the plane. These severely reduced models describe the late-stage evolution of isolated coherent structures in two-dimensional and geostrophic turbulence. For single-layer dynamics, they approximate the relaxation of initially distorted structures toward axisymmetric equilibrium states. For two-layer dynamics, they predict the rate of energy transfer in baroclinically perturbed structures returning to stable barotropic states. Comparisons against direct numerical simulations of the fully-resolved many-vortex dynamics validate the predictive capacity of these reduced models.

Synthesis and formation mechanism of porous silicon carbide stacked by nanoparticles from precipitated silica/glucose composites

Porous silicon carbide (SiC) with a Brunauer–Emmett–Teller specific surface area of 75 m2 g−1 and a pore volume of 0.37 cm3 g−1 was synthesized through a facile process using industrial precipitated silica and glucose. X-ray diffraction, scanning electron microscopy, and transmission electron microscopy analyses indicated that the SiC was stacked by β-SiC nanoparticles with size ranging from 60 to 150 nm, and its morphology was similar to that of precipitated silica. Further analysis showed that the formation mechanism of porous SiC was different from the general carbothermal reduction method. The proposed mechanism indicated that precipitated silica is a porous particle stacked by amorphous primary silica nanoparticles. When the temperature was higher than its softening point during the carbothermal reduction reaction, the silica nanoparticles softened and began to change from solid phase to liquid phase with a certain viscosity. Then, the pyrolytic carbon of glucose in contact with silica nanoparticles was diffused into the liquid phase and reacted in situ to form SiC. Precipitated silica served as a template, and SiC inherited its morphology and structure.

Optimizing adiabatic quantum program compilation using a graph-theoretic framework

Adiabatic quantum computing has evolved in recent years from a theoretical field into an immensely practical area, a change partially sparked by D-Wave System’s quantum annealing hardware. These multimillion-dollar quantum annealers offer the potential to solve optimization problems millions of times faster than classical heuristics, prompting researchers at Google, NASA and Lockheed Martin to study how these computers can be applied to complex real-world problems such as NASA rover missions. Unfortunately, compiling (embedding) an optimization problem into the annealing hardware is itself a difficult optimization problem and a major bottleneck currently preventing widespread adoption. Additionally, while finding a single embedding is difficult, no generalized method is known for tuning embeddings to use minimal hardware resources. To address these barriers, we introduce a graph-theoretic framework for developing structured embedding algorithms. Using this framework, we introduce a biclique virtual hardware layer to provide a simplified interface to the physical hardware. Additionally, we exploit bipartite structure in quantum programs using odd cycle transversal (OCT) decompositions. By coupling an OCT-based embedding algorithm with new, generalized reduction methods, we develop a new baseline for embedding a wide range of optimization problems into fault-free D-Wave annealing hardware. To encourage the reuse and extension of these techniques, we provide an implementation of the framework and embedding algorithms.

Exploiting symmetry for discrete-time reachability computations

We present a method of computing backward reachable sets for nonlinear discrete-time control systems possessing continuous symmetries. The starting point is a dynamic game formulation of reachability analysis where control inputs aim to maintain the state variables within a target tube despite disturbances. Our method exploits symmetry to compute the reachable sets in a lower-dimensional space, enabling a significant computational speedup. To achieve this, we present a general method for symmetry reduction based on the Cartan frame, which simplifies the dynamic programming iteration without algebraic manipulation of the state update equations. We illustrate the results by computing a backward reachable set for a 6-D reach-avoid game of two Dubins vehicles.

Phase reduction and synchronization of a network of coupled dynamical elements exhibiting collective oscillations.

A general phase reduction method for a network of coupled dynamical elements exhibiting collective oscillations, which is applicable to arbitrary networks of heterogeneous dynamical elements, is developed. A set of coupled adjoint equations for phase sensitivity functions, which characterize the phase response of the collective oscillation to small perturbations applied to individual elements, is derived. Using the phase sensitivity functions, collective oscillation of the network under weak perturbation can be described approximately by a one-dimensional phase equation. As an example, mutual synchronization between a pair of collectively oscillating networks of excitable and oscillatory FitzHugh-Nagumo elements with random coupling is studied.

Association study of interaction of tryptophan hydroxylase 2 with serotonin 1A receptor gene polymorphism on major depressive disorder

Objective To explore the interaction between tryptophan hydroxylase 2(TPH2) gene polymorphisms (rs4570625, rs11178997) and serotonin 1A receptor(5-HT1A) gene polymorphisms (rs878567, rs1364043, rs6265) and the association with major depressive disorder (MDD) in a Chinese Han population. Methods The DNA isolated from peripheral blood samples of 288 MDD patients 288 healthy subjects was detected by single base primer extension assay (Snapshot). The generalized multifactor dimensionality reduction (GMDR) method was used to analyze the gene-gene interaction. Results Significant differences were found in the genotype(patients (TT: 27, TA: 152, AA: 109), controls (TT: 82, TA: 105, AA: 101), P<0.01) and allele(patients (T: 206, A: 370), controls (T: 269, A: 307), P<0.01) frequencies of rs11178997 within TPH2 between MDD patients and controls. Statistically, a greater risk of developing MDD was found in individuals with an rs11178997 A-allele(OR=1.574, 95%CI=1.243-1.993). The interaction between TPH2 (rs4570625, rs11178997) and 5-HT1A (rs878567, rs1364043, rs6265) was considered as the best multi-locus model, and this showed a testing accuracy of 57.67% and a CV consistency of 10/10. And this interaction had a significant effect on the risk of MDD (P=0.0107). Conclusion There may be an association between the interaction of TPH2 and 5-HT1A polymorphisms and MDD. Key words: Major depressive disorder; Tryptophan hydroxylase 2; Serotonin 1A receptor; Gene-gene interaction

A General Dimension Reduction Method for the Dispersion Modeling of Semiconductor Devices

This paper presents a general dimension reduction method for the dispersion modeling of current and charge sources of semiconductor devices including HEMTs, LDMOS, and HBTs. The dimensions that are easily handled are represented by the existing empirical or physical functions, while those dimensions that are difficult to tackle with or have limited measurement data are represented by the Taylor expansion. In this paper, an AlGaN/GaN HEMT was taken as the example device to illustrate how this method can be applied to handle the comprehensive dispersion effects induced by thermal and charge trapping. The dimensions of terminal voltages ( $V_{\mathit{ gs}}$ , $V_{\mathit{ ds}}$ ) are characterized by the 10-parameter Angelov function, while the remaining dispersion-related dimensions (channel temperature $T_{j}$ and drain trap state $\phi _{D}$ ) are expressed by the Taylor expansion. The constructed drain current source ( $I_{ds}$ ) model is able to predict a number of pulsed $I$ – $V\text{s}$ with various channel temperature and quiescent biases. Finally, the analytical large signal model was implemented in the advanced design system, and several $RC$ sub-circuits with multiple time constants were exploited to implement the dispersion model in the simulator. Good agreement has been achieved for both small-signal and large-signal characteristics of the investigated devices.

Application of Hilbert-Huang Decomposition to Reduce Noise and Characterize for NMR FID Signal of Proton Precession Magnetometer

The parameters in a nuclear magnetic resonance free induction decay (FID) signal contain information that is useful in magnetic field measurement, magnetic resonance sounding and other related applications. A real time sampled FID signal is well modeled as a finite mixture of exponential sequences plus noise. We propose to use the Hilbert-Huang transform (HHT) for noise reduction and characterization, where the generalized Hilbert-Huang represents a way to decompose a signal into so-called intrinsic mode function (IMF) along with a trend, and obtain instantaneous frequency data. First, acquiring the actual untuned FID signal by a developed prototype of proton magnetometer, and then the empirical mode decomposition is performed to decompose the noise and original FID. Finally, the HHT is applied to the obtained IMFs to extract the Hilbert energy spectrum of the signal on the frequency axis. By theory analysis and the testing of an actual FID signal, the results show that, compared with general noise reduction methods such as auto correlation and singular value decomposition, combined with the proposed method can further suppress the interfered signals effectively, and can obtain different components of FID signal, which can be used to identify the magnetic anomaly, the existence of groundwater etc. This is a very important property since it can be exploited to separate the FID signal from noise and to estimate exponential sequence parameters of FID signal.

CMV reduction for five‐level ANPC converter by PS‐PWM strategy

Five-level active neutral point clamped (ANPC) converter has become one of the most attractive topologies in industry applications, especially in the field of the motor drive. A novel common-mode voltage (CMV) reduction strategy for five-level ANPC is presented, which is based on the phase-shifted pulse-width modulation (PS-PWM). A basic PS-PWM method is presented, which can realise the NP voltage and the flying capacitor (FC) voltages balance. Moreover, a general CMV reduction method is summarised. The proposed CMV reduction strategy can minimise the CMV by injecting zero-sequence voltage to the modulation wave and it does not affect the ability of NP voltage and FC voltages balance. The validity of the proposed CMV reduction method based on PS-PWM strategy is verified by simulation.