Orthogonal Factors Research Articles

Combined Deep Priors With Low-Rank Tensor Factorization for Hyperspectral Image Restoration

Mixed noise pollution severely disturbs hyperspectral image (HSI) processing and applications. Plenty of algorithms have been developed to address this issue via two strategies: model-driven or data-driven strategy. However, model-driven methods exist in the highly time-consuming weakness of iterative optimization and unstable sensitivity of setting parameters. Data-driven methods usually perform poor due to the overfitting effects. To solve these issues, we combine both the deep denoising priors with low-rank tensor factorization (DP-LRTF) for HSI restoration. The proposed method uses Tucker tensor factorization to depict the global spectral low-rank constraint. Then the spectral orthogonal basis and spatial reduced factor are optimized by two deep denoising priors, respectively. Through this integrated strategy, we can simultaneously exploit the intrinsic low-rank property of HSI, and utilize the powerful feature extraction ability by deep learning for HSI restoration. Compared with model-driven and data-driven methods, DP-LRTF outperforms on HSI mixed noise removal and execution efficiency for various simulated/real experiments.

Anchor-Graph Regularized Orthogonal Concept Factorization for Document Clustering

Anchor-Graph Regularized Orthogonal Concept Factorization for Document Clustering

Bayesian hierarchical modeling on covariance valued data

Analysis of structural and functional connectivity (FC) of human brains is of pivotal importance for diagnosis of cognitive ability. The Human Connectome Project (HCP) provides an excellent source of neural data across different regions of interest (ROIs) of the living human brain. Individual specific data were available in the form of time varying covariance matrices representing the brain activity as the subjects perform a specific task. As a preliminary objective of studying the heterogeneity of brain connectomics across the population, we develop a probabilistic model for a sample of covariance matrices using a scaled Wishart distribution. We stress here that our data units are available in the form of covariance matrices, and we use the Wishart distribution to create our likelihood function rather than its more common usage as a prior on covariance matrices. Based on empirical explorations suggesting the data matrices to have a low effective rank, we further model the center of the Wishart distribution using an orthogonal factor model type decomposition. We encourage shrinkage toward a low rank structure through a novel shrinkage prior and discuss strategies to sample from the posterior distribution using a combination of Gibbs and slice sampling. The efficacy of the approach is explored in various simulation settings and exemplified on several case studies including our motivating HCP data. We extend our modeling framework to a dynamic setting to detect change points.

Preparation and performance investigation of a high efficiency cement permeation type waterproofing materials

The active chemicals play the primary waterproof role in Cementitious Capillary Crystalline Waterproofing Materials (CCCW). Based on the combined mechanism of precipitation reaction and complexation precipitation reaction, precipitation component, complexing agent, calcium ion compensator, surfactant, and defoaming agents were determined as the composition of the active chemicals in this study. Based on the multiple performance tests, the material components and the content of active chemicals were investigated with the single and orthogonal factors. The results showed that with the increase of sodium silicate content, the flexural and compressive strengths of the specimens at 28 d increased, and the water absorption of the specimens gradually decreased. When the sodium silicate content was 2 wt%, the flexural and compressive strengths of the specimens at 28 d increased by 55% and 16%, respectively, and the water absorption of the specimens decreased by 27%. When the content of the complexing agent was constant, the setting time of the specimen gradually prolonged with the increase of the proportion of the hydroxyl and carboxyl groups in the molecular structure. Meanwhile, the flexural and compressive strengths of the specimen decreased gradually with the increase of the content of the complexing agent. With the increase of calcium formate content, the strength of the specimens tended to increase first and then decrease. When the ratio between calcium formate and calcium hydroxide was 10:90, the compressive strength of the specimens cured for 28 d increased by 42%. Based on the analysis of the mechanical properties, water absorption, and expansion of the specimen under orthogonal factors, the mixing ratio of the active chemicals was: precipitation component 2 wt%, complexing agent 1 wt%, calcium ion compensator 3 wt%, surfactant 1.5 wt%, and defoaming agents 0.7 wt%. The impermeability pressure ratios of specimens with and without CCCW coating reached to 275% and 200% at 28 d, respectively. The water resistance of the specimens increased with the increase of curing age, and the water absorption of the specimen was reduced by 56% at 28 d.

Behavioural and neural structure of fluent speech production deficits in aphasia.

Deficits in fluent speech production following left hemisphere stroke are a central concern because of their impact on patients' lives and the insight they provide about the neural organization of language processing. Fluent speech production requires the rapid coordination of phonological, semantic, and syntactic processing, so this study examined how deficits in connected speech relate to these language sub-systems. Behavioural data (N = 69 participants with aphasia following left hemisphere stroke) consisted of a diverse and comprehensive set of narrative speech production measures and measures of overall severity, semantic deficits, and phonological deficits. These measures were entered into a principal component analysis with bifactor rotation-a latent structure model where each item loads on a general factor that reflects what is common among the items, and orthogonal factors that explain variance not accounted for by the general factor. Lesion data were available for 58 of the participants, and each factor score was analysed with multivariate lesion-symptom mapping. Effects of connectivity disruption were evaluated using robust regression with tract disconnection or graph theoretic measures of connectivity as predictors. The principal component analysis produced a four-factor solution that accounted for 70.6% of the variance in the data, with a general factor corresponding to the overall severity and length and complexity of speech output (complexity factor), a lexical syntax factor, and independent factors for Semantics and Phonology. Deficits in the complexity of speech output were associated with a large temporo-parietal region, similar to overall aphasia severity. The lexical syntax factor was associated with damage in a relatively small set of fronto-parietal regions which may reflect the recruitment of control systems to support retrieval and correct usage of lexical items that primarily serve a syntactic rather than semantic function. Tract-based measures of connectivity disruption were not statistically associated with the deficit scores after controlling for overall lesion volume. Language network efficiency and average clustering coefficient within the language network were significantly associated with deficit scores after controlling for overall lesion volume. These results highlight overall severity as the critical contributor to fluent speech in post-stroke aphasia, with a dissociable factor corresponding to lexical syntax.

Gaussian Orthogonal Latent Factor Processes for Large Incomplete Matrices of Correlated Data

We introduce Gaussian orthogonal latent factor processes for modeling and predicting large correlated data. To handle the computational challenge, we first decompose the likelihood function of the Gaussian random field with a multi-dimensional input domain into a product of densities at the orthogonal components with lower-dimensional inputs. The continuous-time Kalman filter is implemented to compute the likelihood function efficiently without making approximations. We also show that the posterior distribution of the factor processes is independent, as a consequence of prior independence of factor processes and orthogonal factor loading matrix. For studies with large sample sizes, we propose a flexible way to model the mean, and we derive the marginal posterior distribution to solve identifiability issues in sampling these parameters. Both simulated and real data applications confirm the outstanding performance of this method.

Ultrasonic-Assisted Electrochemical Milling of SLM-Printed Hastelloy X Based on Analysis of Material Passivation Characteristics

Hastelloy X (HX) is widely used in the aerospace field for its excellent corrosion resistance and high-temperature mechanical properties that can be fabricated into complex structures directly by the selective laser melting (SLM) technique. However, SLM-printed (SLM-ed) HX with high strength and hardness is challenging to process using conventional manufacturing techniques and may result in machining flaws that don’t fulfill engineering standards. Therefore, an ultrasonic-assisted electrochemical milling (UAECM) method using a tube electrode is proposed to fabricate high aspect ratio structures on SLM-ed HX with high-quality. Firstly, the passivation characteristics of SLM-ed HX before and after solid solution treatment (SST) were investigated using polarization curves and electrochemical impedance spectroscopy (EIS). Secondly, the electrochemical milling process and the electrode gap flow field were simulated. Then, the effect of processing parameters such as ultrasonic amplitude, electrical parameters, and mechanical parameters on the groove width and stability was investigated by the orthogonal and single factor experiments. Finally, the cavity and bump structures were machined by layered milling with an average groove width of 960 ± 15 μm, a groove depth of 4.4 mm, an aspect ratio of 4.5, and surface roughness of 1.524 μm and 1.622 μm, respectively, demonstrating the adaptability and machining accuracy of the UAECM method.

Central Sensitisation Inventory-Thai version: translation, cross-cultural adaptation, and psychometric evaluation in chronic non-specific neck pain

PurposeTo translate and cross-culturally adapt the English version of the Central Sensitisation Inventory (CSI) into Thai (CSI-Thai) and to evaluate its psychometric properties in individuals with chronic non-specific neck pain.Materials and methodsCross-cultural translation and adaptation of the CSI were performed according to standard guidelines. A total of 340 participants were invited to complete the CSI-Thai, Visual Analogue Scale (pain intensity), Neck Disability Index, Fear-Avoidance Beliefs Questionnaire, Pain Catastrophising Scale (PCS), and Short Form-36. Psychometric evaluation included confirmatory factor analysis, internal consistency, test–retest reliability, agreement, and construct validity.ResultsDimensionality analyses indicated that a bifactor model, comprising one general factor plus four orthogonal factors, fit the CSI structure better than unidimensional and the four-factor models. The general factor showed substantial reliability (Cronbach α = 0.91, Omega ω = 0.94, and omega hierarchical ω–h = 0.91). The intraclass correlation coefficient was 0.90, representing excellent stability over a 48 h interval. Moderate-to-strong correlations and acceptable-to-excellent discriminations were found between the CSI-Thai and all questionnaires. The exception was the PCS (no correlation and discrimination). The standard error of measurement and minimal detectable change of the CSI-Thai were 2.33 and 6.47, respectively.ConclusionsThe translation and cross-cultural adaptation of the CSI-Thai were successful, with satisfactory reliability and construct validity.Implications for rehabilitationCentral Sensitisation Inventory-Thai version (CSI-Thai) is successfully adapted and demonstrated satisfactory reliability and construct validity.The CSI-Thai can be applicable to assess central sensitisation-related signs and symptoms in Thai-speaking patients with chronic non-specific neck pain (CNSNP) both clinical and research purposes.The CSI-Thai correlated to pain, disability and quality of life among patients with CNSNP.

Coequalisers under the Lens

Lenses encode protocols for synchronising systems. We continue the work begun by Chollet et al. at the Applied Category Theory Adjoint School in 2020 to study the properties of the category of small categories and asymmetric delta lenses. The forgetful functor from the category of lenses to the category of functors is already known to reflect monos and epis and preserve epis; we show that it preserves monos, and give a simpler proof that it preserves epis. Together this gives a complete characterisation of the monic and epic lenses in terms of elementary properties of their get functors. Next, we initiate the study of coequalisers of lenses. We observe that not all parallel pairs of lenses have coequalisers, and that the forgetful functor from the category of lenses to the category of functors neither preserves nor reflects all coequalisers. However, some coequalisers are reflected; we study when this occurs, and then use what we learned to show that every epic lens is regular, and that discrete opfibrations have pushouts along monic lenses. Corollaries include that every monic lens is effective, every monic epic lens is an isomorphism, and the class of all epic lenses and the class of all monic lenses form an orthogonal factorisation system.

Steering DNA Condensation on Engineered Nanointerfaces.

DNA has received increasing attention in nanotechnology due to its ability to fold into prescribed structures. Different from the commonly adopted base-pairing strategy, an emerging class of amorphous DNA materials are formed by DNA's abiological interactions. Despite the great successes, a lack of nanoscale nucleation/growth control disables more advanced considerations. This work aims at harnessing the heterogeneous nucleation of metal-ion-glued DNA condensates on nanointerfaces. Upon unveiling key orthogonal factors including solution pH, ionic cross-linkers, and surface functionalities, chemically programmable DNA condensation on nanoparticle seeds is achieved, resembling a famous Stöber process for silica coating. The nucleation rules discovered on individual nanoseeds can be passed on to their dimeric assemblies, where broken spherical symmetry and the existence of interparticle gaps help a regiospecific DNA gelation. The steerable DNA condensation, and the multifunctions from DNA, metal ions, and nanocores, hold a great promise in noncanonical DNA nanotechnology toward novel applications.

Parallel QR Factorization of Block Low-rank Matrices

We present two new algorithms for Householder QR factorization of Block Low-Rank (BLR) matrices: one that performs block-column-wise QR and another that is based on tiled QR. We show how the block-column-wise algorithm exploits BLR structure to achieve arithmetic complexity of 𝒪( mn ), while the tiled BLR-QR exhibits 𝒪( mn 1.5 complexity. However, the tiled BLR-QR has finer task granularity that allows parallel task-based execution on shared memory systems. We compare the block-column-wise BLR-QR using fork-join parallelism with tiled BLR-QR using task-based parallelism. We also compare these two implementations of Householder BLR-QR with a block-column-wise Modified Gram–Schmidt (MGS) BLR-QR using fork-join parallelism and a state-of-the-art vendor-optimized dense Householder QR in Intel MKL. For a matrix of size 131k × 65k, all BLR methods are more than an order of magnitude faster than the dense QR in MKL. Our methods are also robust to ill conditioning and produce better orthogonal factors than the existing MGS-based method. On a CPU with 64 cores, our parallel tiled Householder and block-column-wise Householder algorithms show a speedup of 50 and 37 times, respectively.

Co-consensus semi-supervised multi-view learning with orthogonal non-negative matrix factorization

Semi-supervised multi-view learning has recently achieved appealing performance with the consensus relation between samples. However, in addition to the relation between samples, the relation between samples and their assemble centroid is also important to the learning. In this paper, we propose a novel model based on orthogonal non-negative matrix factorization, which allows exploring both the consensus relations between samples and between samples and their assemble centroid. Since this model utilizes more consensus information to guide the multi-view learning, it can lead to better performance. Meanwhile, we theoretically derive a proposition about the equivalency between the partial orthogonality and the full orthogonality. Based on this proposition, the orthogonality constraint and the label constraint are simultaneously implemented in the proposed model. Experimental evaluations on five real-world datasets show that our approach outperforms the state-of-the-art methods, where the improvement is 6% average in terms of ARI index.

A constrained factor mixture analysis model for consistent and inconsistent respondents to mixed-worded scales.

Mixed-worded scales require a more careful reading and answering process than scales with only one type of wording. The present study is about the unintended consequences of using such scales, because typically, not all respondents answer positively and negatively worded items consistently. This population heterogeneity-meaning that there are distinct groups of consistently and inconsistently answering respondents-may arguably underlie wording-related effects in mixed-worded scales. We formulated a constrained factor mixture analysis model that operationalized these two assumed classes of respondents. We investigated five data sets that contained four mixed-worded attitude scales, large-scale samples from three countries (Germany, Australia, and the U.S.), and two age groups (children and adolescents). The constrained factor mixture analysis model showed estimated parameter patterns in line with theoretical expectations and consistently outperformed its more traditional competitor, confirmatory factor analysis with one global and one orthogonal method factor across all used data sets. We found proportions of between 7% and 20% of respondents belonging to the inconsistent classes. To further substantiate and validate the interpretation of the proposed model, we related class membership to theoretically relevant respondent characteristics such as reading achievement, cognitive skills, or conscientiousness. Further, we undertook an initial exploration of the overlap in inconsistent respondents' class membership across time and across scales within a survey. The article discusses implications for future research as well as for the use of mixed-worded scales. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Parallel solution of optimal control problems using the graphics processing unit

AbstractThis paper presents an indirect method for solving optimal control problems (OCPs) using graphics processing unit (GPU). The OCPs considered here include control variable inequality constraints (CVICs), state variable inequality constraints (SVICs), and parameters. The necessary conditions of the minimum for the OCPs are written as a boundary value problem with index‐1 differential algebraic equations (BVP‐DAEs). The complementarity conditions associated with those inequality constraints are approximated using Kanzow's smoothed Fisher‐Burmeister formula. The numerical solution for solving the BVP‐DAEs is based on the multiple shooting technique and the DAEs are solved using a single step linearly implicit Runge–Kutta (Rosenbrock–Wanner, ROW) method. A Newton's continuation method is performed to solve the BVP‐DAEs system and the descent direction is found by solving a sparse bordered almost block diagonal (BABD) linear system with a structured orthogonal factorization algorithm. Parallel computing techniques are used to accelerate the code which is implemented using Python and CUDA on GPU. Numerical examples are presented to illustrate the efficiency of the code. The GPU based parallel implementation is shown to be significantly faster than the implementation using central processing unit (CPU) alone.

Sensitivity-based constitutive parameter identification of nonlinear structures with unknown input earthquake excitation

System identification methods are widely developed in structural health monitoring, in which the force or structural damage is identified from a linearised structure. However, the real-world civil infrastructures are generically nonlinear and subject to extreme loading conditions, such as earthquake. This study proposes a system identification method for material constitutive nonlinear structures subject to earthquake, where sensitivity analysis is developed to accelerate the complicated nonlinear system identification. First, the nonlinear structural model based on the material constitutional law is constructed by fiber beam-column elements. The earthquake input is parameterized by Chebyshev orthogonal polynomials in a recognizable form. Afterwards, the dynamic response sensitivities with respect to the constitutive parameters and the orthogonal factors of earthquake excitation are derived to accelerate the system identification from a complex material constitutive model. Finally, the constitutive parameters and the earthquake input are simultaneously identified by an optimization process. Numerical structures subjected to earthquake excitation are employed to verify the accuracy of the proposed method, by which the influence of parameter's order, input intensity and measurement noise is discussed. The proposed method develops Chebyshev parameterization to recognize the earthquake input and sensitivity analysis to accelerate the optimization, which leads to an accurate nonlinear system identification of civil structures.

FVS 2.0: A unifying framework for understanding the factors of visual-attentional processing.

Across a broad range of stimulus types and tasks (16 stimulus types × 26 tasks, 1,744 observers in total), the present study employed an individual-item differences analysis to extract the factors of visual-attentional processing. Three orthogonal factors were identified and they can be summarized as an FVS 2.0 framework: featural, visual, and spatial strengths. Apart from one exception (low-level motion), the FVS 2.0 framework accounts for the vast majority (95.4%) of the variances in the 25 tasks. Therefore, the three straightforward factors provide a unifying framework for understanding the relationship between stimulus types as well as those between tasks. Combining these and other related results, the role of preattentive features seems to be rather different from the traditional view: visual features are general purpose, exclusive, innate, constancy based, and keyword like. A general-purpose, exclusive, innate, constancy-based and keyword-like (GEICK) conjecture is proposed which suggests that the features are conscious-level keywords generated by the specific brain area of V4 and/or IT and then used by all other brain areas. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Incomplete multi-view clustering with incomplete graph-regularized orthogonal non-negative matrix factorization

Incomplete multi-view clustering with incomplete graph-regularized orthogonal non-negative matrix factorization

1224-P: Multidimensional Characterization of Prediabetes in the Project Baseline Health Study

Prediabetes is a recognized entity but has been incompletely characterized in a contemporary context. We sought to characterize participants with prediabetes with extensive multi-dimensional clinical and laboratory data and to identify predictors of progression to diabetes. The Project Baseline Health Study (PBHS) is a multi-site prospective U.S. cohort study of 2502 adults with deep clinical phenotyping at enrollment. At enrollment and follow-up, participants were classified by diabetes status (diabetes [DM], prediabetes [preDM], or no diabetes [noDM]) based on glucose, A1c, medications, and self-reported medical history. Principal component analysis (PCA) was performed on 1clinical, survey, and lab variables to create orthogonal factors composed of correlated variables, which were compared between groups using linear models. Logistic regression was used to identify factors associated with progression from preDM to DM. At enrollment, 16participants had noDM; 544 had preDM; and 352 had DM. Over up to 4 years of follow-up, 52 participants with preDM developed DM. PCA identified 33 factors composed of clusters of clinical variables; 24 factors differed significantly by diabetes status: 6 factors significantly differed between noDM and both preDM and DM after adjustment for multiple comparisons, including factors with variables of glucose measures, anthropometry and physical function (q= 1.3 × 10−21) , red blood cell indices (q= 8.3 × 10−10) , lung function (q= 2.0 × 10−6) , cardiac function (q= 0.001) , and risk of chronic disease. Five factors were significantly associated with progression to DM, including factors of anthropometry and physical function [OR (95% CI) 0.6 (0.5,0.9) ], hematologic measures, and cardiac function. Conclusions: PBHS participants with preDM demonstrated pathophysiologic changes in cardiac, pulmonary, and hematology measures and declines in physical function that precede DM and could be markers of higher risk of progression to DM and complications. Disclosure R.Chatterjee montgomery: Research Support; Bristol-Myers Squibb Company, Epigenomics AG. L.C.Kwee: None. N.Pagidipati: Advisory Panel; AstraZeneca, Boehringer Ingelheim International GmbH, Eli Lilly and Company, ESPERION Therapeutics, Inc., Consultant; Boehringer Ingelheim International GmbH, Novo Nordisk, Research Support; Amgen Inc., AstraZeneca, Boehringer Ingelheim International GmbH, Eggland's Best, Eli Lilly and Company, Novartis AG, Novo Nordisk, Sanofi, Verily Life Sciences. A.F.Hernandez: Consultant; Merck & Co., Inc., Novo Nordisk, Research Support; AstraZeneca, Boehringer Ingelheim International GmbH. F.Rodriguez: Advisory Panel; Amgen Inc., Consultant; HealthPals, Novartis Pharmaceuticals Corporation, Novo Nordisk, Stock/Shareholder; Carta Healthcare, Inc. K.W.Mahaffey: Consultant; Novo Nordisk. L.Palaniappan: Consultant; Amgen Inc. S.H.Shah: Research Support; Verily Life Sciences. Funding The Baseline Health Study and this analysis were funded by Verily Life Sciences, San Francisco, California.

Orthogonal multi-view tensor-based learning for clustering

Multi-view spectral clustering aims to improve the performance of spectral clustering through multi-view data. Many multi-view spectral clustering methods have been proposed recently and achieved promising performance. Among these methods, most of them are designed to pursue numerical consistency in multi-view similarity matrices. However, each similarity matrix has its unique statistic distribution, which makes it not appropriate to seek numerical consistency in multi-view similarity matrices or directly average the multi-view similarity matrices. To overcome the aforementioned problem, we propose a novel Orthogonal Multi-view Tensor-based Learning for clustering, abbreviated as OMTL. Specifically, OMTL introduces an orthogonal matrix factorization to eliminate the view-specific statistic distribution and preserve the intrinsic clustering structure of each view, which fully considers the consensus information contained in multiple views to boost multi-view spectral clustering performance. Further, we employ a low-rank tensor constraint to explore the high order correlations among multiple views. By designing an alternating direction method of multipliers (ADMM) based optimization algorithm, the intrinsic similarity matrix of multi-view data can be efficiently learned for spectral clustering. Extensive experiments on several benchmark datasets have illustrated the superior clustering performance of the proposed method compared to several state-of-the-art multi-view clustering methods.

Exploratory and confirmatory analysis to identify airports and airlines characteristics as factors influencing domestic passenger traffic in Nigeria

Passenger traffic at airports is characterised by fluctuations resulting from the influence of several factors. The influence of each factor is different, leading to unpredictable passenger traffic patterns that make planning difficult. Previous studies used geographic, demographic, and economic variables as exploratory factors to examine air travel demand. The study explores several variables to confirm airports and airlines’ characteristics as demand factors for domestic air travel in Nigeria. Data for the study were collected by administering a questionnaire to respondents at major domestic airports in Nigeria. The variables were presented in the 5-point Likert Scale for respondents to rank in order of significance. Exploratory and confirmatory factor analyses (EFA and CFA) were employed to identify the significant factors affecting passenger traffic at domestic airports in Nigeria. EFA reduced fifteen variables to four orthogonal factors influencing passenger traffic at domestic airports. CFA validates airport and airline services, demographics, economic factors, and airport size and facilities as significant factors affecting passenger traffic at domestic airports in Nigeria. The model fit test shows CMIN/DF = 2.263; CFI = 0.940; GFI = 0.929; NFI = 0.901; and RMSEA = 0.078. The result identifies airport and airline characteristics as factors influencing passenger traffic at the domestic airport in any country. It implies that airport and airline characteristics significantly influence domestic air traffic and needs to be included in modelling. Identifying airport and airline characteristics as air travel determinants make this study unique for policy decisions to forecast domestic passenger traffic in a country.