Robust learning of discriminative projection for multicategory classification on the Stiefel manifold

Learning a robust projection with a small number of training samples is still a challenging problem in face recognition, especially when the unseen faces have extreme variation in pose, illumination, and facial expression. To address this problem, we propose a framework formulated under statistical learning theory that facilitates robust learning of a discriminative projection. Dimensionality reduction using the projection matrix is combined with a linear classifier in the regularized framework of lasso regression. The projection matrix in conjunction with the classifier parameters are then found by solving an optimization problem over the Stiefel manifold. The experimental results on standard face databases suggest that the proposed method outperforms some recent regularized techniques when the number of training samples is small.

Face recognition using kernel ridge regression

In this paper, we present novel ridge regression (RR) and kernel ridge regression (KRR) techniques for multivariate labels and apply the methods to the problem of face recognition. Motivated by the fact that the regular simplex vertices are separate points with highest degree of symmetry, we choose such vertices as the targets for the distinct individuals in recognition and apply RR or KRR to map the training face images into a face subspace where the training images from each individual will locate near their individual targets. We identify the new face image by mapping it into this face subspace and comparing its distance to all individual targets. An efficient cross-validation algorithm is also provided for selecting the regularization and kernel parameters. Experiments were conducted on two face databases and the results demonstrate that the proposed algorithm significantly outperforms the three popular linear face recognition techniques (Eigenfaces, Fisherfaces and Laplacianfaces) and also performs comparably with the recently developed Orthogonal Laplacianfaces with the advantage of computational speed. Experimental results also demonstrate that KRR outperforms RR as expected since KRR can utilize the nonlinear structure of the face images. Although we concentrate on face recognition in this paper, the proposed method is general and may be applied for general multi-category classification problems.

Increased cardiometabolic risk is associated with increased TV viewing time

Purpose: Television viewing time, independent of leisure time physical activity, has cross-sectional relationships with the metabolic syndrome and its individual components. We examined whether baseline and 5-yr changes in self-reported television viewing time are associated with changes in continuous biomarkers of cardiometabolic risk (waist circumference, triglycerides, HDL-cholesterol, systolic and diastolic blood pressure, fasting plasma glucose, and a clustered cardiometabolic risk score) in Australian adults.


Methods: The Australian Diabetes, Obesity and Lifestyle Study (AusDiab) is a prospective, population-based cohort study with biological, behavioral, and demographic measures collected in 1999-2000 and 2004-2005. Noninstitutionalized adults aged >=25 yr were measured at baseline (11,247; 55% of those completing an initial household interview); 6400 took part in the 5-yr follow-up biomedical examination, and 3846 met the inclusion criteria for this analysis. Multiple linear regression analysis was used, and unstandardized B coefficients (95% confidence intervals (CI)) are provided.


Results: Baseline television viewing time (10 h·wk-1 unit) was not significantly associated with change in any of the biomarkers of cardiometabolic risk. Increases in television viewing time over 5 yr (10 h·wk-1 unit) were associated with increases in waist circumference (men: 0.43 cm, 95% CI = 0.08-0.78 cm, P = 0.02; women: 0.68 cm, 95% CI = 0.30-1.05, P < 0.001), diastolic blood pressure (women: 0.47 mm Hg, 95% CI = 0.02-0.92 mm Hg, P = 0.04), and the clustered cardiometabolic risk score (women: 0.03, 95% CI = 0.01-0.05, P = 0.007). These associations were independent of baseline television viewing time and baseline and change in physical activity and other potential confounders.


Conclusions: These findings indicate that an increase in television viewing time is associated with adverse cardiometabolic biomarker changes. Further prospective studies using objective measures of several sedentary behaviors are required to confirm causality of the associations found.

Glaucomatous visual field progression with frequency-doubling technology and standard automated perimetry in a longitudinal prospective study

PURPOSE. To compare frequency-doubling technology (FDT) perimetry with standard automated perimetry (SAP) for detecting glaucomatous visual field progression in a longitudinal prospective study.

METHODS. One eye of patients with open-angle glaucoma was tested every 6 months with both FDT and SAP. A minimum of 6 examinations with each perimetric technique was required for inclusion. Visual field progression was determined by two methods: glaucoma change probability (GCP) analysis and linear regression analysis (LRA). For GCP, several criteria for progression were used. The number of locations required to classify progression with FDT compared with SAP, respectively, was 1:2 (least conservative), 1:3, 2:3, 2:4, 2:6, 2:7, 3:6, 3:7, and 3:10 (most conservative). The number of consecutive examinations required to confirm progression was 2-of-3, 2-of-2, and 3-of-3. For LRA, the progression criterion was any significant decline in mean threshold sensitivity over time in each of the following three visual field subdivisions: (1) all test locations, (2) locations in the central 10° and the superior and inferior hemifields, and (3) locations in each quadrant. Using these criteria, the proportion of patients classified as showing progression with each perimetric technique was calculated and, in the case of progression with both, the differences in time to progression were determined.

RESULTS. Sixty-five patients were followed for a median of 3.5 years (median number of examinations, 9). For the least conservative GCP criterion, 32 (49%) patients were found to have progressing visual fields with FDT and 32 (49%) patients with SAP. Only 16 (25%) patients showed progression with both methods, and in most of those patients, FDT identified progression before SAP (median, 12 months earlier). The majority of GCP progression criteria (15/27), classified more patients as showing progression with FDT than with SAP. Contrary to this, more patients showed progression with SAP than FDT, when analysed with LRA; e.g., using quadrant LRA 20 (31%) patients showed progression with FDT, 23 (35%) with SAP, and only 10 (15%) with both.

CONCLUSIONS. FDT perimetry detected glaucomatous visual field progression. However, the proportion of patients who showed progression with both FDT and SAP was small, possibly indicating that the two techniques identify different subgroups of patients. Using GCP, more patients showed progression with FDT than with SAP, yet the opposite occurred using LRA. As there is no independent qualifier of progression, FDT and SAP progression rates vary depending on the method of analysis and the criterion used.

Top-down and bottom-up approaches to motor skill assessment of children : are child-report and parent-report perceptions predictive of children's performance-based assessment results?

Background/ Aim: Therapists use different types of tests, scales, and instruments to assess children's motor skills, including those classified as being top-down and bottom-up. The aim of the study was to investigate the ability of measures of children's motor skill performance from the perspectives of children and parents (a type of top-down assessment) to predict children's performance-based motor ability test results (a type of bottom-up assessment).
Methods: A convenience sample of 38 children and parents was recruited from Victoria, Australia. Motor skill performance was evaluated from a top-down perspective using the Physical Self-Description Questionnaire (PSDQ) and the Movement Assessment Battery for Children – Second Edition (MABC-2) Checklist to measure children's and parents' perspectives respectively. Motor skill performance was also evaluated from a bottom-up approach using the Bruininks-Oseretsky Test of Motor Proficiency – Second Edition (BOT-2). Data were analyzed using multiple linear regression analysis to determine whether the PSDQ or MABC-2 Checklist was predictive of the children's BOT-2 performance results.
Results: Two predictive relationships were identified based on parents' perspectives, where the total score of the MABC-2 Checklist was found to be a significant predictor of the BOT-2 Manual Coordination motor composite score, accounting for 8.35% of its variance, and the BOT-2 Strength and Agility motor composite score, accounting for 11.6% of its variance. No predictive relationships were identified between the children's self-report PSDQ perspectives and the BOT-2 performance scores.
Conclusions: Therapists are encouraged to utilize a combination of top-down and bottom-up approaches and purposefully to seek parents' and children's perspectives when evaluating children's motor skill performance.

The effects of training on performance and performance-related states in individual elite athletes : a dynamic approach

There are difficulties undertaking controlled training studies with elite athletes. Thus, data from non-elite performers are often presented in scientific journals and subsequently used to guide general training principles. This information may not be transferable or specific enough to inform training practices in an individual elite athlete. However, the nature of athletic participation at elite levels provides the opportunity to collect training data, performance-related variables, and performance data of elite athletes over long periods. In this paper, we describe how dynamic linear models provide an opportunity to use these data to inform training. Data from an elite female triathlete collected over a 111-day training period were used to model the relationship between training and self-reported fatigue. The dynamic linear model analysis showed the independent effects of the three modes of triathlon training on fatigue, how these can change across time, and the possible influence of other unmeasured variables. This paper shows the potential for the use of dynamic linear models as an aid to planning training in elite athletes.

A mathematical model for the spread of Strepotococcus pneumoniae with transmission dependent on serotype

We examine a mathematical model for the transmission of Streptococcus Pneumoniae amongst young children when the carriage transmission coefficient depends on the serotype. Carriage means pneumococcal colonization. There are two sequence types (STs) spreading in a population each of which can be expressed as one of two serotypes. We derive the differential equation model for the carriage spread and perform an equilibrium and global stability analysis on it. A key parameter is the effective reproduction number R e. For R e ≤ 1,  there is only the carriage-free equilibrium (CFE) and the carriage will die out whatever be the starting values. For R e > 1, unless the effective reproduction numbers of the two STs are equal, in addition to the CFE there are two carriage equilibria, one for each ST. If the ST with the largest effective reproduction number is initially present, then in the long-term the carriage will tend to the corresponding equilibrium.

A mathematical model for the spread of streptococcus pneumoniae with transmission due to sequence type

This paper discusses a simple mathematical model to describe the spread of Streptococcus pneumoniae. We suppose that the transmission of the bacterium is determined by multi-locus sequence type. The model includes vaccination and is designed to examine what happens in a vaccinated population if MLSTs can exist as both vaccine and non vaccine serotypes with capsular switching possible from the former to the latter. We start off with a discussion of Streptococcus pneumoniae and a review of previous work. We propose a simple mathematical model with two sequence types and then perform an equilibrium and (global) stability analysis on the model. We show that in general there are only three equilibria, the carriage-free equilibrium and two carriage equilibria. If the effective reproduction number Re is less than or equal to one, then the carriage will die out. If Re > 1, then the carriage will tend to the carriage equilibrium corresponding to the multi-locus sequence type with the largest transmission parameter. In the case where both multi-locus sequence types have the same transmission parameter then there is a line of carriage equilibria. Provided that carriage is initially present then as time progresses the carriage will approach a point on this line. The results generalize to many competing sequence types. Simulations with realistic parameter values confirm the analytical results.

Impact of dynamic PHEVs load on renewable sources based distribution system

In this paper, charging effect of dynamic Plug in Hybrid Electric Vehicle (PHEV) is presented in a renewable energy based electricity distribution system. For planning and designing a distribution system, PHEVs are one of the most important factor as it is going to be a spinning reserve of energy, and also a major load for distribution network. A dynamic load model of PHEVs is introduced here based on third order battery model. To determine the system adequacy, it is necessary to do a micro level analysis to know the PHEVs load impact on grid. Scope of such analysis will cover the performance of wind and solar generation with dynamic PHEVs load, as well as the stability analysis of the power grid to demonstrate that it is important to consider the dynamics of PHEVs load in a renewable energy based distribution network.

Predictors of treatment response in young people at ultra-high risk for psychosis who received long-chain omega-3 fatty acids.

Previous efforts in the prospective evaluation of individuals who experience attenuated psychotic symptoms have attempted to isolate mechanisms underlying the onset of full-threshold psychotic illness. In contrast, there has been little research investigating specific predictors of positive outcomes. In this study, we sought to determine biological and clinical factors associated with treatment response, here indexed by functional improvement in a pre-post examination of a 12-week randomized controlled intervention in individuals at ultra-high risk (UHR) for psychosis. Participants received either long-chain omega-3 (ω-3) polyunsaturated fatty acids (PUFAs) or placebo. To allow the determination of factors specifically relevant to each intervention, and to be able to contrast them, both treatment groups were investigated in parallel. Univariate linear regression analysis indicated that higher levels of erythrocyte membrane α-linolenic acid (ALA; the parent fatty acid of the ω-3 family) and more severe negative symptoms at baseline predicted subsequent functional improvement in the treatment group, whereas less severe positive symptoms and lower functioning at baseline were predictive in the placebo group. A multivariate machine learning analysis, known as Gaussian Process Classification (GPC), confirmed that baseline fatty acids predicted response to treatment in the ω-3 PUFA group with high levels of sensitivity, specificity and accuracy. In addition, GPC revealed that baseline fatty acids were predictive in the placebo group. In conclusion, our investigation indicates that UHR patients with higher levels of ALA may specifically benefit from ω-3 PUFA supplementation. In addition, multivariate machine learning analysis suggests that fatty acids could potentially be used to inform prognostic evaluations and treatment decisions at the level of the individual. Notably, multiple statistical analyses were conducted in a relatively small sample, limiting the conclusions that can be drawn from what we believe to be a first-of-its-kind study. Additional studies with larger samples are therefore needed to evaluate the generalizability of these findings.

Formations of robotic swarm: an artificial force based approach

Cooperative control of multiple mobile robots is an attractive and challenging problem which has drawn considerable attention in the recent past. This paper introduces a scalable decentralized control algorithm to navigate a group of mobile robots (swarm) into a predefined shape in 2D space. The proposed architecture uses artificial forces to control mobile agents into the shape and spread them inside the shape while avoiding intermember collisions. The theoretical analysis of the swarm behavior describes the motion of the complete swarm and individual members in relevant situations. We use computer simulated case studies to verify the theoretical assertions and to demonstrate the robustness of the swarm under external disturbances such as death of agents, change of shape etc. Also the performance of the proposed distributed swarm control architecture was investigated in the presence of realistic implementation issues such as localization errors, communication range limitations, boundedness of forces etc.

Women's preferences for selective estrogen reuptake modulators: an investigation using the time trade off technique

PurposeSelective Estrogen Receptor Modulators (SERMs) reduce the risk of breast cancer for women at increased risk by 38%. However, uptake is extremely low and the reasons for this are not completely understood. The aims of this study were to utilize time trade-off methods to determine the degree of risk reduction required to make taking SERMs worthwhile to women, and the factors associated with requiring greater risk reduction to take SERMs. MethodsWomen at increased risk of breast cancer (N = 107) were recruited from two familial cancer clinics in Australia. Participants completed a questionnaire either online or in pen and paper format. Hierarchical multiple linear regression analysis was used to analyze the data. ResultsOverall, there was considerable heterogeneity in the degree of risk reduction required to make taking SERMs worthwhile. Women with higher perceived breast cancer risk and those with stronger intentions to undergo (or who had undergone) an oophorectomy required a smaller degree of risk reduction to consider taking SERMs worthwhile. ConclusionWomen at increased familial risk appear motivated to consider SERMs for prevention. A tailored approach to communicating about medical prevention is essential. Health professionals could usefully highlight the absolute (rather than relative) probability of side effects and take into account an individual’s perceived (rather than objective) risk of breast cancer.

Effect of retinal image defocus on the thickness of the human choroid

PURPOSE: To describe the time-course and amplitude of changes to sub-foveal choroidal thickness (SFCT) induced by imposed hyperopic and myopic retinal defocus and to compare the responses in emmetropic and myopic subjects. METHODS: Twelve East Asian subjects (age: 18-34 years; six were emmetropic and six had myopia between -2.00 and -5.00 dioptres (D)) viewed a distant target (video movie at 6 m) for 60 min on two separate occasions while optical coherence tomography (OCT) images of the choroid were taken in both eyes every 5 min to monitor SFCT. On each occasion, one eye was optimally corrected for distance with a contact lens while the other eye wore a contact lens imposing either 2.00 D hyperopic or 2.00 D myopic retinal defocus. RESULTS: Baseline SFCT in myopic eyes (mean ± S.D.): 256 ± 42 μm was significantly less than in emmetropic eyes (423 ± 62 μm; p < 0.01) and was correlated with magnitude of myopia (-39 μm per dioptre of myopia, R(2) = 0.67: p < 0.01). Repeated measures anova (General Linear Model) analysis revealed that in both subject groups, 2.00 D of myopic defocus caused a rapid increase in SFCT in the defocussed eye (significant by 10 min, increasing to approximately 20 μm within 60 min: p < 0.01), with little change in the control eye. In contrast, 2.00 D of hyperopic defocus caused a decrease in SFCT in the experimental eye (significant by 20-35 min. SFCT decreased by approximately 20 μm within 60 min: p < 0.01) with little change in the control eye. CONCLUSIONS: Small but significant changes in SFCT (5-8%) were caused by retinal defocus. SFCT increased within 10 min of exposure to 2.00 D of monocular myopic defocus, but decreased more slowly in response to 2.00 D of monocular hyperopic defocus. In our relatively small sample we could detect no difference in the magnitude of changes to SFCT caused by defocus in myopic eyes compared to emmetropic eyes.

Physical activity, television viewing time and 12 year changes in waist circumference

PURPOSE: Both moderate-to-vigorous physical activity (MVPA) and sedentary behavior can be associated with adult adiposity. Much of the relevant evidence is from cross-sectional studies or from prospective studies with relevant exposure measures at a single time point prior to weight gain or incident obesity. This study examined whether changes in MVPA and television (TV) viewing time are associated with subsequent changes in waist circumference, using data from three separate observation points in a large population-based prospective study of Australian adults. METHODS: Data were obtained from the Australian Diabetes, Obesity and Lifestyle study collected in 1999-2000 (baseline), 2004-05 (Wave 2), and 2011-12 (Wave 3). The study sample consisted of adults aged 25 to 74 years at baseline who also attended site measurement at three time points (n=3261). Multilevel linear regression analysis examined associations of initial five-year changes in MVPA and TV viewing time (from baseline to Wave 2) with 12-year change in waist circumference (from baseline to Wave 3), adjusting for well-known confounders. RESULTS: As categorical predictors, increases in MVPA significantly attenuated increases in waist circumference (p for trend< 0.001). TV viewing time change was not significantly associated with changes in waist circumference (p for trend =0.06). Combined categories of MVPA and TV viewing time changes were predictive of waist circumference increases; compared to those who increased MVPA and reduced TV viewing time, those who reduced MVPA and increased TV viewing time had a 2cm greater increase in waist circumference (p=0.001). CONCLUSION: Decreasing MVPA emerged as a significant predictor of increases in waist circumference. Increasing TV viewing time was also influential, but its impact was much weaker than MVPA.This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially.