scholarly journals Robustness Can Be Cheap: A Highly Efficient Approach to Discover Outliers under High Outlier Ratios

2019 ◽  
Vol 33 ◽  
pp. 5313-5320
Author(s):  
Siqi Wang ◽  
En Zhu ◽  
Xiping Hu ◽  
Xinwang Liu ◽  
Qiang Liu ◽  
...  
Keyword(s):  
Computational Cost ◽  
Solution Process ◽  
Closed Form Solution ◽  
Form Solution ◽  
Low Rank ◽  
Alternative Formulation ◽  
Efficient Detection ◽  
Comparable Performance ◽  
High Space ◽  
High Outlier

Efficient detection of outliers from massive data with a high outlier ratio is challenging but not explicitly discussed yet. In such a case, existing methods either suffer from poor robustness or require expensive computations. This paper proposes a Low-rank based Efficient Outlier Detection (LEOD) framework to achieve favorable robustness against high outlier ratios with much cheaper computations. Specifically, it is worth highlighting the following aspects of LEOD: (1) Our framework exploits the low-rank structure embedded in the similarity matrix and considers inliers/outliers equally based on this low-rank structure, which facilitates us to encourage satisfying robustness with low computational cost later; (2) A novel re-weighting algorithm is derived as a new general solution to the constrained eigenvalue problem, which is a major bottleneck for the optimization process. Instead of the high space and time complexity (O((2n)2)/O((2n)3)) required by the classic solution, our algorithm enjoys O(n) space complexity and a faster optimization speed in the experiments; (3) A new alternative formulation is proposed for further acceleration of the solution process, where a cheap closed-form solution can be obtained. Experiments show that LEOD achieves strong robustness under an outlier ratio from 20% to 60%, while it is at most 100 times more memory efficient and 1000 times faster than its previous counterpart that attains comparable performance. The codes of LEOD are publicly available at https://github.com/demonzyj56/LEOD.

scholarly journals X-RAY CT Reconstruction by using Spatially Non Homogeneous ICD Optimization

2019 ◽  
Vol 8 (6S3) ◽  
pp. 2043-2046
Keyword(s):  
Search Algorithm ◽  
Computational Cost ◽  
Selection Criterion ◽  
Closed Form Solution ◽  
Helical Ct ◽  
Reconstruction Algorithm ◽  
Form Solution ◽  
Ct Reconstruction ◽  
Practical Applications ◽  
Voxel Selection

Recent applications of conventional iterative coordinate descent (ICD) algorithms to multislice helical CT reconstructions have shown that conventional ICD can greatly improve image quality by increasing resolution as well as reducing noise and some artifacts. However, high computational cost and long reconstruction times remain as a barrier to the use of conventional algorithm in the practical applications. Among the various iterative methods that have been studied for conventional, ICD has been found to have relatively low overall computational requirements due to its fast convergence. This paper presents a fast model-based iterative reconstruction algorithm using spatially nonhomogeneous ICD (NH-ICD) optimization. The NH-ICD algorithm speeds up convergence by focusing computation where it is most needed. The NH-ICD algorithm has a mechanism that adaptively selects voxels for update. First, a voxel selection criterion VSC determines the voxels in greatest need of update. Then a voxel selection algorithm VSA selects the order of successive voxel updates based upon the need for repeated updates of some locations, while retaining characteristics for global convergence. In order to speed up each voxel update, we also propose a fast 3-D optimization algorithm that uses a quadratic substitute function to upper bound the local 3-D objective function, so that a closed form solution can be obtained rather than using a computationally expensive line search algorithm. The experimental results show that the proposed method accelerates the reconstructions by roughly a factor of three on average for typical 3-D multislice geometries.

Toward a Minimal Dynamic Model of a 6-DOF Parallel Robot

1997 ◽  
Author(s):  
L. Beji ◽  
M. Pascal ◽  
P. Joli
Keyword(s):  
Dynamic Model ◽  
Degrees Of Freedom ◽  
Equations Of Motion ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Parallel Robot ◽  
Form Solution ◽  
Lagrangian Formalism ◽  
Inverse Kinematic Problem ◽  
Geometrical Properties

Abstract In this paper, an architecture of a six degrees of freedom (dof) parallel robot and three limbs is described. The robot is called Space Manipulator (SM). In a first step, the inverse kinematic problem for the robot is solved in closed form solution. Further, we need to inverse only a 3 × 3 passive jacobian matrix to solve the direct kinematic problem. In a second step, the dynamic equations are derived by using the Lagrangian formalism where the coordinates are the passive and active joint coordinates. Based on geometrical properties of the robot, the equations of motion are derived in terms of only 9 coordinates related by 3 kinematic constraints. The computational cost of the obtained dynamic model is reduced by using a minimum set of base inertial parameters.

scholarly journals A Block-Based Regularized Approach for Image Interpolation

2014 ◽  
Vol 2014 ◽  
pp. 1-8
Author(s):  
Li Chen ◽  
Xiaotong Huang ◽  
Jing Tian
Keyword(s):  
Cost Function ◽  
Regularization Parameter ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Image Interpolation ◽  
Adaptive Strategy ◽  
Form Solution ◽  
Interpolation Techniques ◽  
Block Based ◽  
Value Decomposition

This paper presents a new efficient algorithm for image interpolation based on regularization theory. To render ahigh-resolution(HR) image from alow-resolution(LR) image, classical interpolation techniques estimate the missing pixels from the surrounding pixels based on a pixel-by-pixel basis. In contrast, the proposed approach formulates the interpolation problem into the optimization of a cost function. The proposed cost function consists of a data fidelity term and regularization functional. The closed-form solution to the optimization problem is derived using the framework of constrained least squares minimization, by incorporating Kronecker product andsingular value decomposition(SVD) to reduce the computational cost of the algorithm. The effect of regularization on the interpolation results is analyzed, and an adaptive strategy is proposed for selecting the regularization parameter. Experimental results show that the proposed approach is able to reconstruct high-fidelity HR images, while suppressing artifacts such as edge distortion and blurring, to produce superior interpolation results to that of conventional image interpolation techniques.

scholarly journals Hybrid TOA/RSS Range-Based Localization with Self-Calibration in Asynchronous Wireless Networks

2019 ◽  
Vol 8 (2) ◽  
pp. 31 ◽  
Author(s):  
Angelo Coluccia ◽  
Alessio Fascista
Keyword(s):  
Low Cost ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Target Position ◽  
Form Solution ◽  
Position Estimation ◽  
Model Parameters ◽  
Time Of Arrival ◽  
Synchronization Errors ◽  
Step Algorithm

The paper addresses the problem of localization based on hybrid received signal strength (RSS) and time of arrival (TOA) measurements, in the presence of synchronization errors among all the nodes in a wireless network, and assuming all parameters are unknown. In most existing schemes, in fact, knowledge of the model parameters is postulated to reduce the high dimensionality of the cost functions involved in the position estimation process. However, such parameters depend on the operational wireless context, and change over time due to the presence of dynamic obstacles and other modification of the environment. Therefore, they should be adaptively estimated “on the field”, with a procedure that must be as simple as possible in order to suit multiple real-time re-calibrations, even in low-cost applications, without requiring human intervention. Unfortunately, the joint maximum likelihood (ML) position estimator for this problem does not admit a closed-form solution, and numerical optimization is practically unfeasible due to the large number of nuisance parameters. To circumvent such issues, a novel two-step algorithm with reduced complexity is proposed: A first calibration phase exploits nodes in known positions to estimate the unknown RSS and TOA model parameters; then, in a second localization step, an hybrid TOA/RSS range estimator is combined with an iterative least-squares procedure to finally estimate the unknown target position. The results show that the proposed hybrid TOA/RSS localization approach outperformed state-of-the-art competitors and, remarkably, achieved almost the same accuracy of the joint ML benchmark but with a significantly lower computational cost.

A Closed Form Solution of Dual-Phase Lag Heat Conduction Problem With Time Periodic Boundary Conditions

2019 ◽  
Vol 141 (3) ◽  
Author(s):  
Pranay Biswas ◽  
Suneet Singh ◽  
Hitesh Bindra
Keyword(s):  
Heat Conduction ◽  
Boundary Conditions ◽  
Closed Form ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Form Solution ◽  
Fourier Series Expansion ◽  
Phase Lag ◽  
Dual Phase ◽  
Time Periodic

The Laplace transform (LT) is a widely used methodology for analytical solutions of dual phase lag (DPL) heat conduction problems with consistent DPL boundary conditions (BCs). However, the inversion of LT requires a series summation with large number of terms for reasonably converged solution, thereby, increasing computational cost. In this work, an alternative approach is proposed for this inversion which is valid only for time-periodic BCs. In this approach, an approximate convolution integral is used to get an analytical closed-form solution for sinusoidal BCs (which is obviously free of numerical inversion or series summation). The ease of implementation and simplicity of the proposed alternative LT approach is demonstrated through illustrative examples for different kind of sinusoidal BCs. It is noted that the solution has very small error only during the very short initial transient and is (almost) exact for longer time. Moreover, it is seen from the illustrative examples that for high frequency periodic BCs the Fourier and DPL model give quite different results; however, for low frequency BCs the results are almost identical. For nonsinusoidal periodic function as BCs, Fourier series expansion of the function in time can be obtained and then present approach can be used for each term of the series. An illustrative example with a triangular periodic wave as one of the BC is solved and the error with different number of terms in the expansion is shown. It is observed that quite accurate solutions can be obtained with a fewer number of terms.

Exact solutions for the macro-, meso- and micro-scale analysis of composite laminates and sandwich structures

2018 ◽  
Vol 52 (22) ◽  
pp. 3109-3124 ◽  
Author(s):  
Yang Yan ◽  
Alfonso Pagani ◽  
Erasmo Carrera ◽  
Qingwen Ren
Keyword(s):  
Composite Laminates ◽  
Computational Cost ◽  
Three Dimensional ◽  
Sandwich Beam ◽  
Closed Form Solution ◽  
Single Layer ◽  
Transversely Isotropic ◽  
Global Scale ◽  
Form Solution ◽  
Lagrange Polynomials

The present work proposes a closed-form solution based on refined beam theories for the static analysis of fiber-reinforced composite and sandwich beams under simply supported boundary conditions. The higher-order beam models are developed by employing Carrera Unified Formulation, which uses Lagrange-polynomials expansions to approximate the kinematic field over the cross section. The proposed methodology allows to carry out analysis of composite structure analysis through a single formulation in global-local sense, i.e. homogenized laminates at a global scale and fiber-matrix constituents at a local scale, leading to component-wise analysis. Therefore, three-dimensional stress/displacement fields at different scales can be successfully detected by increasing the order of Lagrange polynomials opportunely. The governing equations are derived in a strong-form and solved in a Navier-type sense. Three benchmark numerical assessments are carried out on a single-layer transversely isotropic beam, a cross-ply laminate [Formula: see text] beam and a sandwich beam. The results show that accurate displacement and stress values can be obtained in different parts of the structure with lower computational cost in comparison with traditional, enhanced as well as three-dimensional finite element methods. Besides, this study may serve as benchmarks for future assessments in this field.

scholarly journals An analytic solution of full-sky spherical geometry for satellite relative motions

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Soung Sub Lee ◽  
Christopher D. Hall
Keyword(s):  
Relative Motion ◽  
Analytic Solution ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Spherical Geometry ◽  
Form Solution ◽  
Geometrical Approach ◽  
Satellite Relative Motion ◽  
Modeling Accuracy ◽  
Spherical Triangles

AbstractHerein, an exact and efficient analytic solution for an unperturbed satellite relative motion was developed using a direct geometrical approach. The derivation of the relative motion geometrically interpreted the projected Keplerian orbits of the satellites on a sphere (Earth and celestial spheres) using the solutions of full-sky spherical triangles. The results were basic and computationally faster than the vector and plane geometry solutions owing to the advantages of the full-sky spherical geometry. Accordingly, the validity of the proposed solution was evaluated by comparing it with other analytic relative motion theories in terms of modeling accuracy and efficiency. The modeling accuracy showed an equivalent performance with Vadali’s nonlinear unit sphere approach, which is essentially equal to the Yan–Alfriend nonlinear theory. Moreover, the efficiency was demonstrated by the lowest computational cost compared with other models. In conclusion, the proposed modeling approach illustrates a compact and efficient closed-form solution for satellite relative motion dynamics.

scholarly journals Efficient Inexact Proximal Gradient Algorithm for Nonconvex Problems

2017 ◽  
Author(s):  
Quanming Yao ◽  
James T. Kwok ◽  
Fei Gao ◽  
Wei Chen ◽  
Tie-Yan Liu
Keyword(s):  
Matrix Completion ◽  
Closed Form Solution ◽  
Image Inpainting ◽  
Form Solution ◽  
The State ◽  
Gradient Algorithm ◽  
Nonconvex Problems ◽  
Proximal Gradient Algorithm ◽  
Comparable Performance ◽  
Completion Problems

While proximal gradient algorithm is originally designed for convex optimization, several variants have been recently proposed for nonconvex problems. Among them, nmAPG [Li and Lin, 2015] is the state-of-art. However, it is inefficient when the proximal step does not have closed-form solution, or such solution exists but is expensive, as it requires more than one proximal steps to be exactly solved in each iteration. In this paper, we propose an efficient accelerate proximal gradient (niAPG) algorithm for nonconvex problems. In each iteration, it requires only one inexact (less expensive) proximal step. Convergence to a critical point is still guaranteed, and a O(1/k) convergence rate is derived. Experiments on image inpainting and matrix completion problems demonstrate that the proposed algorithm has comparable performance as the state-of-the-art, but is much faster.

scholarly journals STAP Based on Toeplitz Covariance Matrix Reconstruction for Airborne Radar

2020 ◽  
Vol 2020 ◽  
pp. 1-10
Author(s):  
Mingxin Liu ◽  
Lin Zou ◽  
Xuelian Yu ◽  
Yun Zhou ◽  
Xuegang Wang
Keyword(s):  
Covariance Matrix ◽  
Degrees Of Freedom ◽  
Coupling Effect ◽  
Closed Form Solution ◽  
Linear Structure ◽  
Form Solution ◽  
Small Sample ◽  
Low Rank ◽  
Matrix Reconstruction ◽  
Rank Matrix

We consider the problem of clutter covariance matrix (CCM) estimation for space-time adaptive processing (STAP) radar in the small sample. In this paper, a fast efficient algorithm for CCM reconstruction is proposed to overcome this shortcoming for the linear structure. Particularly, we present a low-rank matrix recovery (LRMR) question about CCM estimation based on the Toeplitz structure of CCM and the prior knowledge of the noise. The closed-form solution is obtained by relaxing the nonconvex LRMR problem that the trace norm replaces the rank norm. The target can then be efficiently detected by using the recovered CCM according to the STAP theorem. We also analyze the algorithm model under the linear structure in the presence of unknown mutual coupling. It is shown that our method can obtain accurate CCM in the small sample, with even higher accuracy than the traditional algorithms in the same number of samples. It also can reduce the coupling effect and obtain more degrees of freedom (DOF) with limited sensors and pulses by utilizing sparse linear structure (SLS) and improve angle and Doppler resolutions. Finally, numerical simulations have verified the effectiveness of the proposed method in comparison with some of the existing methods.

scholarly journals An Efficient and Robust Unsupervised Anomaly Detection Method Using Ensemble Random Projection in Surveillance Videos

Sensors ◽  
2019 ◽  
Vol 19 (19) ◽  
pp. 4145 ◽  
Author(s):  
Jingtao Hu ◽  
En Zhu ◽  
Siqi Wang ◽  
Xinwang Liu ◽  
Xifeng Guo ◽  
...  
Keyword(s):  
Anomaly Detection ◽  
Detection Method ◽  
Computational Cost ◽  
Closed Form Solution ◽  
Random Projection ◽  
Form Solution ◽  
Least Square ◽  
Support Vector ◽  
Regression Problem ◽  
Unsupervised Anomaly Detection

Video anomaly detection is widely applied in modern society, which is achieved by sensors such as surveillance cameras. This paper learns anomalies by exploiting videos under the fully unsupervised setting. To avoid massive computation caused by back-prorogation in existing methods, we propose a novel efficient three-stage unsupervised anomaly detection method. In the first stage, we adopt random projection instead of autoencoder or its variants in previous works. Then we formulate the optimization goal as a least-square regression problem which has a closed-form solution, leading to less computational cost. The discriminative reconstruction losses of normal and abnormal events encourage us to roughly estimate normality that can be further sifted in the second stage with one-class support vector machine. In the third stage, to eliminate the instability caused by random parameter initializations, ensemble technology is performed to combine multiple anomaly detectors’ scores. To the best of our knowledge, it is the first time that unsupervised ensemble technology is introduced to video anomaly detection research. As demonstrated by the experimental results on several video anomaly detection benchmark datasets, our algorithm robustly surpasses the recent unsupervised methods and performs even better than some supervised approaches. In addition, we achieve comparable performance contrast with the state-of-the-art unsupervised method with much less running time, indicating the effectiveness, efficiency, and robustness of our proposed approach.

Sign in / Sign up

Export Citation Format

Share Document