Complete Algorithm Research Articles

Sound source localization of non-synchronous measurements beamforming based on the truncated nuclear norm regularization

Due to the size and aperture limitations of the microphone array, poor resolution and limited working frequency range are challenging for acoustic imaging with beamforming methods. In this work, two fast iteration algorithms, alternating direction method of multipliers (ADMM) and accelerated proximal gradient line search method (APGL) based on truncated nuclear norm regularization (TNNR), are proposed to complete the cross-spectral matrix for non-synchronous measurements beamforming. Compared with current matrix completion algorithm based on nuclear norm minimization (NNM), TNNR is regarded as a better approximation to the rank function, which means it can approach the real solution better and restore the sound sources accurately at a low signal-to-noise ratio (SNR) environment. Among the TNNR-APGL and TNNR-ADMM, the TNNR-APGL algorithm has a lower matrix completion error and a shorter calculation time. The localization results of TNNR-ADMM method are more robust to the number of measurements. The experiment results show that TNNR methods greatly improve the resolution of conventional beamforming (CBF) and suppress the sidelobes under low SNR conditions, which indicates that the non-synchronous measurements beamforming based on TNNR methods has the potential to be applied in the industrial scene and low SNR environment.

Complete Neighbourhood Search Heuristic Algorithm for Portfolio Optimization

In portfolio optimization, the fundamental goal of an investor is to optimally allocate investments between different assets. Mean-variance optimization methods make unrealistic assumptions to solve the problem of optimal allocation. On the other hand, when realistic constraints like holding size and cardinality are introduced it leads to optimal asset allocation which differ from the mean variance optimization. The resulting optimization problem become quite complex as it exhibits multiple local extrema and discontinuities. Heuristic algorithms work well for the complex problem. Therefore, a heuristic algorithm is developed which is based on hill climbing complete (HC-C). It is utilized to solve the extended portfolio optimization problem. In order to validate its performance, the proposed HC-C is tested with standard portfolio optimization problem. Experimental results are benchmarked with the quadratic programming method and threshold accepting (TA) algorithm.

Quantum algorithm for neighborhood preserving embedding

Neighborhood preserving embedding (NPE) is an important linear dimensionality reduction technique that aims at preserving the local manifold structure. NPE contains three steps, i.e., finding the nearest neighbors of each data point, constructing the weight matrix, and obtaining the transformation matrix. Liang et al. proposed a variational quantum algorithm (VQA) for NPE [Phys. Rev. A 101 032323 (2020)]. The algorithm consists of three quantum sub-algorithms, corresponding to the three steps of NPE, and was expected to have an exponential speedup on the dimensionality n. However, the algorithm has two disadvantages: (i) It is not known how to efficiently obtain the input of the third sub-algorithm from the output of the second one. (ii) Its complexity cannot be rigorously analyzed because the third sub-algorithm in it is a VQA. In this paper, we propose a complete quantum algorithm for NPE, in which we redesign the three sub-algorithms and give a rigorous complexity analysis. It is shown that our algorithm can achieve a polynomial speedup on the number of data points m and an exponential speedup on the dimensionality n under certain conditions over the classical NPE algorithm, and achieve a significant speedup compared to Liang et al.’s algorithm even without considering the complexity of the VQA.

A generalized Mallows model based on [formula omitted]-divergence measures

Rankings occur when a set of items is ordered in agreement with some criteria or personal opinions and can be found in various problems ranging from voting and elections to food preferences. Distances on permutations are commonly used in rank data analysis and are an efficient tool for constructing probability models for rankings. In this paper, we consider the optimal property of the distance-based Mallows model in terms of the Kullback–Leibler divergence and propose a generalization based on the ϕ-divergence. In the sequel, we focus on a special parametric family of models induced by the Cressie–Read power divergence. For the suggested models, we provide parameter estimating algorithms and model fitting methods. Furthermore, we propose a simple approach for the specification of the consensus ranking, based on modal complete or partial rankings modal rankings, that could be used alternatively to complete search algorithms. As an illustration, the described procedures are applied to three examples of rank data.

A beam search algorithm for minimizing crane times in premarshalling problems

The premarshalling problem consists of sorting the containers placed in a bay of the container yard so that they can be retrieved in the order in which they will be required. We study the premarshalling problem with crane time minimization objective and develop a beam search algorithm, with some new elements adapted to the characteristics of the problem, to solve it. We propose various evaluation criteria, depending on the type of container movement, for its local evaluation; a new heuristic algorithm including local search for blue its global evaluation; and several new dominance rules. The computational study shows the contribution of each new element. The performance of the complete algorithm is tested on well-known benchmarks. The beam search algorithm matches all known optimal solutions, improves on the known suboptimal solutions, and obtains solutions for the largest instances, for which no solution had previously been found.

Dynamic Controllability of Processes without Surprises

Dynamic controllability guarantees that a process control can steer the execution of a business process without violating any temporal constraints although some tasks have uncontrollable durations. However, it has been shown that dynamic controllability may lead to process executions with the undesirable property that tasks have to be started or ended on extremely short notice. Sudden termination forces the agent to immediately terminate the execution of a process task without any prior notice in order to meet some temporal constraints. Semi-dynamic controllability guarantees dynamic controllability and the absence of sudden termination. Here, we show that dynamic controllability may also lead to the problem of sudden start, which forces the immediate start of a process task without prior notice. We formalize all constellations of temporal constraints causing sudden start and sudden termination in a process. We propose a technique to design processes in which activities can be dynamically dispatched without these surprises, i.e., with advance notice, and extend the notion of semi-dynamic controllability by also considering the sudden start. This leads to a sound and complete algorithm for checking the semi-dynamic controllability of time-constrained processes.

Illinois-Type Methods for Noisy Euclidean Distance Realization

In this work, we introduce an iterative algorithm for the Euclidean distance matrix completion (EDMC) problem with noisy and incomplete distance measurements. The proposed method is based on semidefinite programming, utilizes a Pareto iterative approach, and performs a projection-free convex optimization over the spectrahedron to solve a level-set problem relevant to EDMC problems. The optimality trade-off between the trace of a positive semidefinite matrix and a loss function is pursued over Pareto optimal points with simple, derivative-free, costly efficient nonlinear equation root finding iterations called Illinois-type methods. We evaluate our approach numerically in a scenario where distance measurements are affected by multiplicative noise.

Outsourcing LDA-Based Face Recognition to an Untrusted Cloud

Face recognition has been extensively employed in practice, such as attendance system and public security. Linear discriminant analysis (LDA) algorithm is one of the most significant ones in the field of face recognition, but it is very difficult for many clients to employ it in their resource-constrained devices (e.g., smartphones and notebook computers).Outsourcing computation provides a promising method for clients to perform heavy tasks with limited computing power. In this paper, we design a protocol of outsourcing LDA-based face recognition to an untrusted cloud, which can help the client to complete the operations of matrix inversion (MI), matrix multiplication (MM) and eigenvalue decomposition (ED) simultaneously. The proposed outsourcing protocol can hide the private data of the client from the cloud server. More importantly, the client can verify whether the outsourcing results are correct or not with probability one and so it is impossible for the server to deceive the client. In addition, the proposed protocol greatly decreases the computational complexity of the client thus enabling the client to complete LDA algorithm efficiently. Finally, we implement the protocol and give a comprehensive evaluation. The experimental results demonstrate that the client obtain great computing savings and the face recognition accuracy in the proposed protocol is almost identical to the original LDA algorithm.

DLSLA 3-D SAR Imaging via Sparse Recovery Through Combination of Nuclear Norm and Low-Rank Matrix Factorization

Downward-looking sparse linear array 3-D synthetic aperture radar (DLSLA 3-D SAR) cross-track dimensional imaging always suffers from incomplete observation which does not satisfy the Nyquist sampling theorem and leads to the failure of conventional 3-D frequency-domain methods. Although several sparse reconstruction-based methods have been presented to solve this problem, the basis mismatch issue in sparse reconstruction theory will degrade the image reconstruction performance. To address this issue, this article proposes a novel 3-D imaging method for DLSLA 3-D SAR, which provides another idea for 3-D imaging through sparse recovery. It utilizes recovered full-sampled data to achieve cross-track dimensional imaging instead of using the under-sampled data directly as before. The Along-track-Height plane imaging is first finished by the range-Doppler (RD) algorithm and motion error compensation. Then, an advanced nuclear norm and low-rank matrix factorization (NU-LRMF)-based matrix completion (MC) algorithm and a vector reconstruction framework are built to achieve accurate recovery of full-sampled data. Finally, the cross-track dimensional imaging is completed with recovered full-sampled data by geometric correction and beamforming. Moreover, a fast two-stage iteration strategy for NU-LRMF (TS-NU-LRMF) is also presented to accelerate convergence. The robustness and effectiveness of the proposed 3-D imaging method are verified by several numerical simulations and comparative studies based on both the complex 3-D ship model and the simulated 3-D distributed scenario.

Optimizing Daylight Glare and Circadian Entrainment in a Daylight-Artificial Light Integrated Scheme

Human circadian rhythm varies with different combinations of light sources. Natural light and dark patterns regulate the sleeping and waking cycles of the human body. This study aimed to develop a simplified model for daylight glare and circadian entrainment in an existing controlled daylight-artificial light integrated system. In addition, the effects of window orientation, occupant position, artificial light, daylight, and a combination of both on Circadian Stimulus (CS) were investigated. In a daylit interior space, window orientation, shading control and occupant seating position play a significant role in the CS. Simplified models to predict Daylight Glare Probability (DGP) and CS were estimated from the dataset. The developed nonlinear CS estimation models have R-squared values of 0.983 and 0.974. The precision of the models was evaluated in terms of standard error of regression (S) and AIC score. Multi-Objective Genetic Algorithm based optimization technique was developed using the above models to minimize glare and maximize CS in the working space. This optimisation methodology aids in implementing a complete automation algorithm for the test workbench to achieve visual comfort and circadian entrainment.

Spatiotemporal Urban Inference and Prediction in Sparse Mobile CrowdSensing: a Graph Neural Network Approach

Mobile CrowdSensing (MCS) has recently become a promising data acquisition paradigm, which recruits a large number of users to collect data from the target sensing areas. Obviously, with the increase of sensing scale and the decrease of sensing granularity, traditional MCS cannot fully cover the required sensing areas especially the inaccessible areas. As a variant, Sparse MCS can utilize the spatiotemporal correlations in sensing data to infer the whole sensing map only by sensing a few subareas. However, in many real-world scenarios, such as traffic congestion prediction or parking occupancy detection, inferring the current unsensed data may not be the final goal. By comparison, it is more important to get the future information through the sparse sensed data. In this paper, we turn attention from inferring the current unsensed data to predicting the future unknown data and propose an urban inference and prediction framework in Sparse MCS. To deal with the sparse sensed data, we first present a bipartite-graph-based matrix completion algorithm with spatiotemporal constraints to accurately recover the current full map. Then, by exploiting spatiotemporal correlations based on the inferred full map, we present a Graph Convolutional Networks (GCN) with spatiotemporal attention to predict the future maps. Furthermore, we design a spatiotemporal iterative method to repeatedly update the spatiotemporal attentions and constraints, in order to connect the urban inference and prediction to improve the accuracy of the whole framework. Extensive experiments have been conducted on two types of typical urban sensing tasks with four real-world data sets, which verify the effectiveness of our proposed algorithms in improving the inference and prediction accuracy with the sparse sensed data.

Mc2g: An Efficient Algorithm for Matrix Completion With Social and Item Similarity Graphs

In this paper, we design and analyze MC2G (Matrix Completion with 2 Graphs), an algorithm that performs matrix completion in the presence of social and item similarity graphs. MC2G runs in quasilinear time and is parameter free. It is based on spectral clustering and local refinement steps. For the matrix completion problem which possesses additional block structures in its rows and columns, we derive the expected number of sampled entries required for MC2G to succeed, and further show that it matches an information-theoretic lower bound up to a constant factor for a wide range of parameters. We perform extensive experiments on both synthetic datasets and a semi-real dataset inspired by real graphs. The experimental results show that MC2G outperforms other state-of-the-art matrix completion algorithms.

Grassmannian Optimization for Online Tensor Completion and Tracking With the t-SVD

We propose a new fast streaming algorithm for the tensor completion problem of imputing missing entries of a low-tubal-rank tensor using the tensor singular value decomposition (t-SVD) algebraic framework. We show the t-SVD is a specialization of the well-studied block-term decomposition for third-order tensors, and we present an algorithm under this model that can track changing free submodules from incomplete streaming 2-D data. The proposed algorithm uses principles from incremental gradient descent on the Grassmann manifold of subspaces to solve the tensor completion problem with linear complexity and constant memory in the number of time samples. We provide a local expected linear convergence result for our algorithm. Our empirical results are competitive in accuracy but much faster in compute time than state-of-the-art tensor completion algorithms on real applications to recover temporal chemo-sensing and MRI data under limited sampling.

Deep Unrolled Low-Rank Tensor Completion for High Dynamic Range Imaging.

The major challenge in high dynamic range (HDR) imaging for dynamic scenes is suppressing ghosting artifacts caused by large object motions or poor exposures. Whereas recent deep learning-based approaches have shown significant synthesis performance, interpretation and analysis of their behaviors are difficult and their performance is affected by the diversity of training data. In contrast, traditional model-based approaches yield inferior synthesis performance to learning-based algorithms despite their theoretical thoroughness. In this paper, we propose an algorithm unrolling approach to ghost-free HDR image synthesis algorithm that unrolls an iterative low-rank tensor completion algorithm into deep neural networks to take advantage of the merits of both learning- and model-based approaches while overcoming their weaknesses. First, we formulate ghost-free HDR image synthesis as a low-rank tensor completion problem by assuming the low-rank structure of the tensor constructed from low dynamic range (LDR) images and linear dependency among LDR images. We also define two regularization functions to compensate for modeling inaccuracy by extracting hidden model information. Then, we solve the problem efficiently using an iterative optimization algorithm by reformulating it into a series of subproblems. Finally, we unroll the iterative algorithm into a series of blocks corresponding to each iteration, in which the optimization variables are updated by rigorous closed-form solutions and the regularizers are updated by learned deep neural networks. Experimental results on different datasets show that the proposed algorithm provides better HDR image synthesis performance with superior robustness compared with state-of-the-art algorithms, while using significantly fewer training samples.

Two-stage Characteristic Recognition of Demand Side Resource for Load Aggregator Based on Bias-SVD Matrix Factorization and Bayesian Inference

As the agent and risk taker of demand side resources, inaccurate recognition of user characteristics will lead to large trial-and-error cost of load aggregator (LA) at the early stage of market-oriented demand response (DR). In this work, the typical invitation-based DR program implemented in China's provincial DR trading pilot is first introduced, and a bidding and scheduling strategy optimization model is built as a simulation tool for LA's decision-making in the DR transactions. Then, a two-stage user characteristic recognition method is proposed for LA to improve its market strategies based on a labelled demand side resources database named demand resource pool (DRP). In the first stage of the method, the DRP data completion algorithm based on Bias- singular value decomposition matrix factorization is proposed to preliminarily recognize the potential demand side resources with unknown characteristic labels. In the second stage, the Bayesian inference-based modification algorithm is proposed to realize a refined recognition of user characteristic labels according to LA's empirical conclusions and the new samples of each round of the DR transactions. Simulations based on the data from the emerging DR pilot in Zhejiang province, China illustrate that the proposed method has lower error in user characteristic recognition, which is conducive to LA to bear lower loss risk and enhance its income at the early stage of DR trading.

Potential of vision-enhanced floating car data for urban traffic estimation

Floating car data (FCD) have recently become a popular tool of urban traffic engineering. Conventional FCD contains a series of probe cars’ timestamped locations and are used to estimate traffic speeds and travel times and identify congestions. In this study, we propose the enhancement of conventional FCD with car vision information: traffic measurements, collected by video cameras or lidars, installed on the probe cars. Given the limited penetration rates of probe cars, such vision information can significantly improve the accuracy of traffic estimation. Our experimental study is based on two simulated vision-enhanced FCD data sets: sensor-based real-world traffic data set with simulated observability and a simulation model with vision-equipped probe cars. We estimate the potential of vision-enhanced FCD for urban traffic flow estimation for different probe car penetration rates. A recently proposed temporal geometric matrix completion algorithm is utilized for traffic speed estimation given incomplete spatiotemporal traffic flow information. Empirical results show that the availability of vision-enhanced FCD leads to significant improvement: the coverage of 6% of spatiotemporal slots by probe cars gives reasonable, and the coverage of 9% – nearly optimal accuracy of traffic speed estimation. Thus, obtained experimental results support our hypothesis about the utility of vision-enhanced FCD to improve traffic estimation accuracy and discover related problems and limitations.

Color Image Recovery Using Low-Rank Quaternion Matrix Completion Algorithm.

As a new color image representation tool, quaternion has achieved excellent results in color image processing problems. In this paper, we propose a novel low-rank quaternion matrix completion algorithm to recover missing data of a color image. Motivated by two kinds of low-rank approximation approaches (low-rank decomposition and nuclear norm minimization) in traditional matrix-based methods, we combine the two approaches in our quaternion matrix-based model. Furthermore, the nuclear norm of the quaternion matrix is replaced by the sum of the Frobenius norm of its two low-rank factor quaternion matrices. Based on the relationship between the quaternion matrix and its equivalent complex matrix, the problem eventually is converted from the quaternion number domain to the complex number domain. An alternating minimization method is applied to solve the model. Simulation results on color image recovery show the superior performance and efficiency of the proposed algorithm over some tensor-based and quaternion-based ones.

An extensive solution to prevent wax deposition formation in gas-lift wells

In recent years, an insurmountable problem in high-wax oil production has been the intensive formation of organic deposits in the bottom-hole zone of the formation, downhole and surface equipment, leading to a dramatic loss in wells production, transportation capability, and subsequently has a detrimental effect on efficiency. Gas-lift is a widely used method, where compressed high-pressure gas is injected through the tubing annulus. Consequently, the current fluid level in the annulus decreases and increases in the tubing string. The compressed gas enters the tubing through gas-lift valves, mixing with the liquid. During production, changes in pressure, temperature, and fluid composition along the well bore highly depend upon the compressed gas injected into the well. At temperatures below the wax appearance temperature, organic wax crystals will form either in bulk volumes of fluid or on cold surfaces of equipment, consequently, covering the surface. Wax formation is a hazardous and costly task in the oil and gas industry, especially when operating gas-lift wells. Hence, the prevention of wax formation in gas-lift wells is of utmost importance. Although, numerous methods have been implemented to handle the wax formation problem in recent decades, questions concerning their economic viability as well as their effectiveness remain relevant. In this paper, a complete algorithm for determining changes in the oil component composition in a gas-lift well during the injection of associated petroleum gas is detailed. The obtained results play a pivotal role in the study of wax deposition. A technological scheme has been developed to prevent wax deposit formation in the tubing strings during gas-lift well operation. A simulation using the introduced process and technology has been run, and an observed result obtained. After running the simulation, we determined the optimal composition of the associated petroleum injection gas based on the required flow rate of the working agent and the lowest value of the wax appearance temperature. This method offers potential solutions to increasing gas-lift well production efficiency under complex wax formation conditions.

Propagation Map Reconstruction via Interpolation Assisted Matrix Completion

Constructing a propagation map from a set of scattered measurements finds important applications in many areas, such as localization, spectrum monitoring and management. Classical interpolation-type methods have poor performance in regions with very sparse measurements. Recent advance in matrix completion has the potential to reconstruct a propagation map from sparse measurements, but the spatial resolution is limited. This paper proposes to integrate interpolation with matrix completion to exploit both the spatial correlation and the potential low rank structure of the propagation map. The proposed method first enriches matrix observations using interpolation, and develops the statistics of the interpolation error based on a local polynomial regression model. Then, two uncertainty-aware matrix completion algorithms are developed to exploit the interpolation error statistics. It is numerically demonstrated that the proposed method outperforms Kriging and other state-of-the-art schemes, and reduces the mean squared error (MSE) of propagation map reconstruction by 10%-50% for a medium to large number of measurements.

Inter operator variability of machine learning researchers predicting all-cause mortality in patients admitted to intensive care unit

Abstract Background The number of publications using machine learning (ML) to predict cardiovascular outcomes and identify clusters of patients at greater risk has risen dramatically in recent years. However, research papers which use ML often fail to provide sufficient information about their algorithms to enable results to be replicated by others in the same or different datasets. Aim To test the reproducibility of results from ML algorithms given three different levels of information commonly found in publications: model type alone, a description of the model, and complete algorithm. Methods MIMIC-III is a healthcare dataset comprising detailed information from over 60,000 intensive care unit (ICU) admissions from the Beth Israel Deaconess Medical Centre between 2001 and 2012. Access is available to everyone pending approval and completion of a short training course. Using this dataset, three models for predicting all-cause in-hospital mortality were created, two from a PhD student working in ML, and one from an existing research paper which used the same dataset and provided complete model information. A second researcher (a PhD student in ML and cardiology) was given the same dataset and was tasked with reproducing their results. Initially, this second researcher was told what type of model was created in each case, followed by a brief description of the algorithms. Finally, the complete algorithms from each participant were provided. In all three scenarios, recreated models were compared to original models using Area Under the Receiver Operating Characteristic Curve (AUC). Results After excluding those younger than 18 years and events with missing or invalid entries, 21,139 ICU admissions remained from 18,094 patients between 2001 and 2012, including 2,797 in-hospital deaths. Three models were produced: two Recurrent Neural Networks (RNNs) which differed significantly in internal weights and variables, and a Boosted Tree Classifier (BTC). The AUC of the first reproduced RNN matched that of the original RNN (Figure 1), however the second RNN and the BTC could not be reproduced given model type alone. As more information was provided about these algorithms, the results from the reproduced models matched the original results more closely. Conclusions In order to create clinically useful ML tools with results that are reproducible and consistent, it is vital that researchers share enough detail about their models. Model type alone is not enough to guarantee reproducibility. Although some models can be recreated with limited information, this is not always the case, and the best results are found when the complete algorithm is shared. These findings have huge relevance when trying to apply ML in clinical practice. Funding Acknowledgement Type of funding sources: None.