Applying SVMs and weight-based factor analysis to unsupervised adaptation for speaker verification

This paper presents an extended study on the implementation of support vector machine(SVM) based speaker verification in systems that employ continuous progressive model adaptation using the weight-based factor analysis model. The weight-based factor analysis model compensates for session variations in unsupervised scenarios by incorporating trial confidence measures in the general statistics used in the inter-session variability modelling process. Employing weight-based factor analysis in Gaussian mixture models (GMM) was recently found to provide significant performance gains to unsupervised classification. Further improvements in performance were found through the integration of SVM-based classification in the system by means of GMM supervectors. This study focuses particularly on the way in which a client is represented in the SVM kernel space using single and multiple target supervectors. Experimental results indicate that training client SVMs using a single target supervector maximises performance while exhibiting a certain robustness to the inclusion of impostor training data in the model. Furthermore, the inclusion of low-scoring target trials in the adaptation process is investigated where they were found to significantly aid performance.

Effects of different practice task constraints on fluctuations of player heart rate in small-sided football games

This paper analyzes effects of different practice task constraints on heart rate (HR) variability during 4v4 smallsided football games. Participants were sixteen football players divided into two age groups (U13, Mean age: 12.4±0.5 yrs; U15: 14.6±0.5). The task consisted of a 4v4 sub-phase without goalkeepers, on a 25x15 m field, of 15 minutes duration with an active recovery period of 6 minutes between each condition. We recorded players’ heart rates using heart rate monitors (Polar Team System, Polar Electro, Kempele, Finland) as scoring mode was manipulated (line goal: scoring by dribbling past an extended line; double goal: scoring in either of two lateral goals; and central goal: scoring only in one goal). Subsequently, %HR reserve was calculated with the Karvonen formula. We performed a time-series analysis of HR for each individual in each condition. Mean data for intra-participant variability showed that autocorrelation function was associated with more short-range dependence processes in the “line goal” condition, compared to other conditions, demonstrating that the “line goal” constraint induced more randomness in HR response. Relative to inter-individual variability, line goal constraints demonstrated lower %CV and %RMSD (U13: 9% and 19%; U15: 10% and 19%) compared with double goal (U13: 12% and 21%; U15: 12% and 21%) and central goal (U13: 14% and 24%; U15: 13% and 24%) task constraints, respectively. Results suggested that line goal constraints imposed more randomness on cardiovascular stimulation of each individual and lower inter-individual variability than double goal and central goal constraints.

The ecological dynamics of 1v1 sub-phases in association football

This paper analyzes effects of different practice task constraints on heart rate (HR) variability during 4v4 smallsided football games. Participants were sixteen football players divided into two age groups (U13, Mean age: 12.4±0.5 yrs; U15: 14.6±0.5). The task consisted of a 4v4 sub-phase without goalkeepers, on a 25x15 m field, of 15 minutes duration with an active recovery period of 6 minutes between each condition. We recorded players’ heart rates using heart rate monitors (Polar Team System, Polar Electro, Kempele, Finland) as scoring mode was manipulated (line goal: scoring by dribbling past an extended line; double goal: scoring in either of two lateral goals; and central goal: scoring only in one goal). Subsequently, %HR reserve was calculated with the Karvonen formula. We performed a time-series analysis of HR for each individual in each condition. Mean data for intra-participant variability showed that autocorrelation function was associated with more short-range dependence processes in the “line goal” condition, compared to other conditions, demonstrating that the “line goal” constraint induced more randomness in HR response. Relative to inter-individual variability, line goal constraints demonstrated lower %CV and %RMSD (U13: 9% and 19%; U15: 10% and 19%) compared with double goal (U13: 12% and 21%; U15: 12% and 21%) and central goal (U13: 14% and 24%; U15: 13% and 24%) task constraints, respectively. Results suggested that line goal constraints imposed more randomness on cardiovascular stimulation of each individual and lower inter-individual variability than double goal and central goal constraints.

Development and validation of a pregnancy symptoms inventory

BACKGROUND. Physical symptoms are common in pregnancy and are predominantly associated with normal physiological changes. These symptoms have a social and economic cost, leading to absenteeism from work and additional medical interventions. There is currently no simple method for identifying common pregnancy related problems in the antenatal period. A validated tool, for use by pregnancy care providers would be useful. AIM: The aim of the project was to develop and validate a Pregnancy Symptoms Inventory for use by healthcare professionals (HCPs). METHODS: A list of symptoms was generated via expert consultation with midwives and obstetrician gynaecologists. Focus groups were conducted with pregnant women in their first, second or third trimester. The inventory was then tested for face validity and piloted for readability and comprehension. For test-re-test reliability, it was administered to the same women 2 to 3 days apart. Finally, outpatient midwives trialled the inventory for 1 month and rated its usefulness on a 10cm visual analogue scale (VAS). The number of referrals to other health care professionals was recorded during this month. RESULTS: Expert consultation and focus group discussions led to the generation of a 41-item inventory. Following face validity and readability testing, several items were modified. Individual item test re-test reliability was between .51 to 1 with the majority (34 items) scoring .0.70. During the testing phase, 211 surveys were collected in the 1 month trial. Tiredness (45.5%), poor sleep (27.5%) back pain (19.5%) and nausea (12.6%) were experienced often. Among the women surveyed, 16.2% claimed to sometimes or often be incontinent. Referrals to the incontinence nurse increased > 8 fold during the study period. The median rating by midwives of the ‘usefulness’ of the inventory was 8.4 (range 0.9 to 10). CONCLUSIONS: The Pregnancy Symptoms Inventory (PSI) was well accepted by women in the 1 month trial and may be a useful tool for pregnancy care providers and aids clinicians in early detection and subsequent treatment of symptoms. It shows promise for use in the research community for assessing the impact of lifestyle intervention in pregnancy.

HyperCard test engine using Rasch Analysis

This paper will focus on the development of an interactive test engine using Rasch analysis of item responses for question selection and reporting of results. The Rasch analysis is used to determine student ability and question difficulty. This model is widely used in the preparation of paper-based tests and has been the subject of particular use and development at the Australian Council for Education Research (ACER). This paper presents an overview of an interactive implementation of the Rasch analysis model in HyperCard, where student ability estimates are generated 'on the fly' and question difficulty values updated from time to time. The student ability estimates are used to determine question selection and are the basis of scoring and reporting schemes.

The 'nurdle to leg' and other ways of winning cricket matches

The emergence of Twenty20 cricket at the elite level has been marketed on the excitement of the big hitter, where it seems that winning is a result of the muscular batter hitting boundaries at will. This version of the game has captured the imagination of many young players who all want to score runs with “big hits”. However, in junior cricket, boundary hitting is often more difficult due to size limitations of children and games played on outfields where the ball does not travel quickly. As a result, winning is often achieved via a less spectacular route – by scoring more singles than your opponents. However, most standard coaching texts only describe how to play boundary scoring shots (e.g. the drives, pulls, cuts and sweeps) and defensive shots to protect the wicket. Learning to bat appears to have been reduced to extremes of force production, i.e. maximal force production to hit boundaries or minimal force production to stop the ball from hitting the wicket. Initially, this is not a problem because the typical innings of a young player (<12 years) would be based on the concept of “block” or “bash” – they “block” the good balls and “bash” the short balls. This approach works because there are many opportunities to hit boundaries off the numerous inaccurate deliveries of novice bowlers. Most runs are scored behind the wicket by using the pace of the bowler’s delivery to re-direct the ball, because the intrinsic dynamics (i.e. lack of strength) of most children means that they can only create sufficient power by playing shots where the whole body can contribute to force production. This method works well until the novice player comes up against more accurate bowling when they find they have no way of scoring runs. Once batters begin to face “good” bowlers, batters have to learn to score runs via singles. In cricket coaching manuals (e.g. ECB, n.d), running between the wickets is treated as a separate task to batting, and the “basics” of running, such as how to “back- up”, carry the bat, calling and turning and sliding the bat into the crease are “drilled” into players. This task decomposition strategy focussing on techniques is a common approach to skill acquisition in many highly traditional sports, typified in cricket by activities where players hit balls off tees and receive “throw-downs” from coaches. However, the relative usefulness of these approaches in the acquisition of sporting skills is increasingly being questioned (Pinder, Renshaw & Davids, 2009). We will discuss why this is the case in the next section.

Prediction of cis/trans isomerization in proteins using PSI-BLAST profiles and secondary structure information

Background The majority of peptide bonds in proteins are found to occur in the trans conformation. However, for proline residues, a considerable fraction of Prolyl peptide bonds adopt the cis form. Proline cis/trans isomerization is known to play a critical role in protein folding, splicing, cell signaling and transmembrane active transport. Accurate prediction of proline cis/trans isomerization in proteins would have many important applications towards the understanding of protein structure and function. Results In this paper, we propose a new approach to predict the proline cis/trans isomerization in proteins using support vector machine (SVM). The preliminary results indicated that using Radial Basis Function (RBF) kernels could lead to better prediction performance than that of polynomial and linear kernel functions. We used single sequence information of different local window sizes, amino acid compositions of different local sequences, multiple sequence alignment obtained from PSI-BLAST and the secondary structure information predicted by PSIPRED. We explored these different sequence encoding schemes in order to investigate their effects on the prediction performance. The training and testing of this approach was performed on a newly enlarged dataset of 2424 non-homologous proteins determined by X-Ray diffraction method using 5-fold cross-validation. Selecting the window size 11 provided the best performance for determining the proline cis/trans isomerization based on the single amino acid sequence. It was found that using multiple sequence alignments in the form of PSI-BLAST profiles could significantly improve the prediction performance, the prediction accuracy increased from 62.8% with single sequence to 69.8% and Matthews Correlation Coefficient (MCC) improved from 0.26 with single local sequence to 0.40. Furthermore, if coupled with the predicted secondary structure information by PSIPRED, our method yielded a prediction accuracy of 71.5% and MCC of 0.43, 9% and 0.17 higher than the accuracy achieved based on the singe sequence information, respectively. Conclusion A new method has been developed to predict the proline cis/trans isomerization in proteins based on support vector machine, which used the single amino acid sequence with different local window sizes, the amino acid compositions of local sequence flanking centered proline residues, the position-specific scoring matrices (PSSMs) extracted by PSI-BLAST and the predicted secondary structures generated by PSIPRED. The successful application of SVM approach in this study reinforced that SVM is a powerful tool in predicting proline cis/trans isomerization in proteins and biological sequence analysis.

Effective attacking strategies in rugby union

Evasive change-of-direction manoeuvres (agility skills) are a fundamental ability in rugby union. In this study, we explored the attributes of agility skill execution as they relate to effective attacking strategies in rugby union. Seven Super 14 games were coded using variables that assessed team patterns and individual movement characteristics during attacking ball carries. The results indicated that tackle-breaks are a key determinant of try-scoring ability and team success in rugby union. The ability of the attacking ball carrier to receive the ball at high speed with at least two body lengths from the defence line against an isolated defender promoted tackle-breaks. Furthermore, the execution of a side-step evasive manoeuvre at a change of direction angle of 20–60° and a distance of one to two body lengths from the defence, and then straightening the running line following the initial direction change at an angle of 20–60°, was associated with tackle-breaks. This study provides critical insight regarding the attributes of agility skill execution that are associated with effective ball carries in rugby union.

Oxygen delivery through high-flow nasal cannulae increase end-expiratory lung volume and reduce respiratory rate in post-cardiac surgical patients

Background: High-flow nasal cannulae (HFNC) create positive oropharyngeal airway pressure but it is unclear how their use affects lung volume. Electrical impedance tomography (EIT) allows assessment of changes in lung volume by measuring changes in lung impedance. Primary objectives were to investigate the effects of HFNC on airway pressure (Paw) and end-expiratory lung volume (EELV), and to identify any correlation between the two. Secondary objectives were to investigate the effects of HFNC on respiratory rate (RR), dyspnoea, tidal volume and oxygenation; and the interaction between body mass index (BMI) and EELV. Methods: Twenty patients prescribed HFNC post-cardiac surgery were investigated. Impedance measures, Paw, PaO2/FiO2 ratio, RR and modified Borg scores were recorded first on low flow oxygen (nasal cannula or Hudson face mask) and then on HFNC. Results: A strong and significant correlation existed between Paw and end-expiratory lung impedance (EELI) (r=0.7, p<0.001). Compared with low flow oxygen, HFNC significantly increased EELI by 25.6% (95% CI 24.3, 26.9) and Paw by 3.0 cmH2O (95% CI 2.4, 3.7). RR reduced by 3.4 breaths per minute (95% CI 1.7, 5.2) with HFNC use, tidal impedance variation increased by 10.5% (95% CI 6.1, 18.3) and PaO2/FiO2 ratio improved by 30.6 mmHg (95% CI 17.9, 43.3). HFNC improved subjective dyspnoea scoring (p=0.023). Increases in EELI were significantly influenced by BMI, with larger increases associated with higher BMIs (p<0.001). Conclusions: This study suggests that HFNC improve dyspnoea and oxygenation by increasing both EELV and tidal volume, and are most beneficial in patients with higher BMIs.

Multifractal characterisation and analysis of complex networks

Complex networks have been studied extensively due to their relevance to many real-world systems such as the world-wide web, the internet, biological and social systems. During the past two decades, studies of such networks in different fields have produced many significant results concerning their structures, topological properties, and dynamics. Three well-known properties of complex networks are scale-free degree distribution, small-world effect and self-similarity. The search for additional meaningful properties and the relationships among these properties is an active area of current research. This thesis investigates a newer aspect of complex networks, namely their multifractality, which is an extension of the concept of selfsimilarity. The first part of the thesis aims to confirm that the study of properties of complex networks can be expanded to a wider field including more complex weighted networks. Those real networks that have been shown to possess the self-similarity property in the existing literature are all unweighted networks. We use the proteinprotein interaction (PPI) networks as a key example to show that their weighted networks inherit the self-similarity from the original unweighted networks. Firstly, we confirm that the random sequential box-covering algorithm is an effective tool to compute the fractal dimension of complex networks. This is demonstrated on the Homo sapiens and E. coli PPI networks as well as their skeletons. Our results verify that the fractal dimension of the skeleton is smaller than that of the original network due to the shortest distance between nodes is larger in the skeleton, hence for a fixed box-size more boxes will be needed to cover the skeleton. Then we adopt the iterative scoring method to generate weighted PPI networks of five species, namely Homo sapiens, E. coli, yeast, C. elegans and Arabidopsis Thaliana. By using the random sequential box-covering algorithm, we calculate the fractal dimensions for both the original unweighted PPI networks and the generated weighted networks. The results show that self-similarity is still present in generated weighted PPI networks. This implication will be useful for our treatment of the networks in the third part of the thesis. The second part of the thesis aims to explore the multifractal behavior of different complex networks. Fractals such as the Cantor set, the Koch curve and the Sierspinski gasket are homogeneous since these fractals consist of a geometrical figure which repeats on an ever-reduced scale. Fractal analysis is a useful method for their study. However, real-world fractals are not homogeneous; there is rarely an identical motif repeated on all scales. Their singularity may vary on different subsets; implying that these objects are multifractal. Multifractal analysis is a useful way to systematically characterize the spatial heterogeneity of both theoretical and experimental fractal patterns. However, the tools for multifractal analysis of objects in Euclidean space are not suitable for complex networks. In this thesis, we propose a new box covering algorithm for multifractal analysis of complex networks. This algorithm is demonstrated in the computation of the generalized fractal dimensions of some theoretical networks, namely scale-free networks, small-world networks, random networks, and a kind of real networks, namely PPI networks of different species. Our main finding is the existence of multifractality in scale-free networks and PPI networks, while the multifractal behaviour is not confirmed for small-world networks and random networks. As another application, we generate gene interactions networks for patients and healthy people using the correlation coefficients between microarrays of different genes. Our results confirm the existence of multifractality in gene interactions networks. This multifractal analysis then provides a potentially useful tool for gene clustering and identification. The third part of the thesis aims to investigate the topological properties of networks constructed from time series. Characterizing complicated dynamics from time series is a fundamental problem of continuing interest in a wide variety of fields. Recent works indicate that complex network theory can be a powerful tool to analyse time series. Many existing methods for transforming time series into complex networks share a common feature: they define the connectivity of a complex network by the mutual proximity of different parts (e.g., individual states, state vectors, or cycles) of a single trajectory. In this thesis, we propose a new method to construct networks of time series: we define nodes by vectors of a certain length in the time series, and weight of edges between any two nodes by the Euclidean distance between the corresponding two vectors. We apply this method to build networks for fractional Brownian motions, whose long-range dependence is characterised by their Hurst exponent. We verify the validity of this method by showing that time series with stronger correlation, hence larger Hurst exponent, tend to have smaller fractal dimension, hence smoother sample paths. We then construct networks via the technique of horizontal visibility graph (HVG), which has been widely used recently. We confirm a known linear relationship between the Hurst exponent of fractional Brownian motion and the fractal dimension of the corresponding HVG network. In the first application, we apply our newly developed box-covering algorithm to calculate the generalized fractal dimensions of the HVG networks of fractional Brownian motions as well as those for binomial cascades and five bacterial genomes. The results confirm the monoscaling of fractional Brownian motion and the multifractality of the rest. As an additional application, we discuss the resilience of networks constructed from time series via two different approaches: visibility graph and horizontal visibility graph. Our finding is that the degree distribution of VG networks of fractional Brownian motions is scale-free (i.e., having a power law) meaning that one needs to destroy a large percentage of nodes before the network collapses into isolated parts; while for HVG networks of fractional Brownian motions, the degree distribution has exponential tails, implying that HVG networks would not survive the same kind of attack.

Useful field of view predicts driving in the presence of distracters

Purpose. The Useful Field of View (UFOV(R)) test has been shown to be highly effective in predicting crash risk among older adults. An important question which we examined in this study is whether this association is due to the ability of the UFOV to predict difficulties in attention-demanding driving situations that involve either visual or auditory distracters. Methods. Participants included 92 community-living adults (mean age 73.6 +/- 5.4 years; range 65-88 years) who completed all three subtests of the UFOV involving assessment of visual processing speed (subtest 1), divided attention (subtest 2), and selective attention (subtest 3); driving safety risk was also classified using the UFOV scoring system. Driving performance was assessed separately on a closed-road circuit while driving under three conditions: no distracters, visual distracters, and auditory distracters. Driving outcome measures included road sign recognition, hazard detection, gap perception, time to complete the course, and performance on the distracter tasks. Results. Those rated as safe on the UFOV (safety rating categories 1 and 2), as well as those responding faster than the recommended cut-off on the selective attention subtest (350 msec), performed significantly better in terms of overall driving performance and also experienced less interference from distracters. Of the three UFOV subtests, the selective attention subtest best predicted overall driving performance in the presence of distracters. Conclusions. Older adults who were rated as higher risk on the UFOV, particularly on the selective attention subtest, demonstrated poorest driving performance in the presence of distracters. This finding suggests that the selective attention subtest of the UFOV may be differentially more effective in predicting driving difficulties in situations of divided attention which are commonly associated with crashes.

A 2-h diagnostic protocol to assess patients with chest pain symptoms in the Asia-Pacific region (ASPECT): a prospective observational validation study

Background: Patients with chest pain contribute substantially to emergency department attendances, lengthy hospital stay, and inpatient admissions. A reliable, reproducible, and fast process to identify patients presenting with chest pain who have a low short-term risk of a major adverse cardiac event is needed to facilitate early discharge. We aimed to prospectively validate the safety of a predefined 2-h accelerated diagnostic protocol (ADP) to assess patients presenting to the emergency department with chest pain symptoms suggestive of acute coronary syndrome. Methods: This observational study was undertaken in 14 emergency departments in nine countries in the Asia-Pacific region, in patients aged 18 years and older with at least 5 min of chest pain. The ADP included use of a structured pre-test probability scoring method (Thrombolysis in Myocardial Infarction [TIMI] score), electrocardiograph, and point-of-care biomarker panel of troponin, creatine kinase MB, and myoglobin. The primary endpoint was major adverse cardiac events within 30 days after initial presentation (including initial hospital attendance). This trial is registered with the Australia-New Zealand Clinical Trials Registry, number ACTRN12609000283279. Findings: 3582 consecutive patients were recruited and completed 30-day follow-up. 421 (11•8%) patients had a major adverse cardiac event. The ADP classified 352 (9•8%) patients as low risk and potentially suitable for early discharge. A major adverse cardiac event occurred in three (0•9%) of these patients, giving the ADP a sensitivity of 99•3% (95% CI 97•9–99•8), a negative predictive value of 99•1% (97•3–99•8), and a specificity of 11•0% (10•0–12•2). Interpretation: This novel ADP identifies patients at very low risk of a short-term major adverse cardiac event who might be suitable for early discharge. Such an approach could be used to decrease the overall observation periods and admissions for chest pain. The components needed for the implementation of this strategy are widely available. The ADP has the potential to affect health-service delivery worldwide.

Non-destructive evaluation of articular cartilage defects using near-infrared (NIR) spectroscopy in osteoarthritic rat models and its direct relation to Mankin Score

Objective The aim of this study was to demonstrate the potential of near-infrared (NIR) spectroscopy for categorizing cartilage degeneration induced in animal models. Method Three models of osteoarthritic degeneration were induced in laboratory rats via one of the following methods: (i) menisectomy (MSX); (ii) anterior cruciate ligament transaction (ACLT); and (iii) intra-articular injection of mono-ido-acetete (1 mg) (MIA), in the right knee joint, with 12 rats per model group. After 8 weeks, the animals were sacrificed and tibial knee joints were collected. A custom-made nearinfrared (NIR) probe of diameter 5 mm was placed on the cartilage surface and spectral data were acquired from each specimen in the wavenumber range 4 000 – 12 500 cm−1. Following spectral data acquisition, the specimens were fixed and Safranin–O staining was performed to assess disease severity based on the Mankin scoring system. Using multivariate statistical analysis based on principal component analysis and partial least squares regression, the spectral data were then related to the Mankinscores of the samples tested. Results Mild to severe degenerative cartilage changes were observed in the subject animals. The ACLT models showed mild cartilage degeneration, MSX models moderate, and MIA severe cartilage degenerative changes both morphologically and histologically. Our result demonstrate that NIR spectroscopic information is capable of separating the cartilage samples into different groups relative to the severity of degeneration, with NIR correlating significantly with their Mankinscore (R2 = 88.85%). Conclusion We conclude that NIR is a viable tool for evaluating articularcartilage health and physical properties such as change in thickness with degeneration.

Interactive conference voting

We describe and analyze opinion polling results from interactive voting procedures undertaken before and after presentations during the Outcome Measures in Rheumatoid Arthritis Clinical Trials Conference (OMERACT II) in Ottawa, Canada, June 30-July 2, 1994. The scoring procedure was a matched voting design; when a participant used the same keypad at the beginning and end of voting, change within a participant could be estimated. Participants, experienced in the rheumatic diseases included clinicians, researchers, methodologists, regulators, and representatives of the pharmaceutical industry. Patients under consideration were those with any rheumatic diseases. Questions were constructed to evaluate the change in voting behavior expected from the content of the presentation. Statistically significant and substantively important changes were evident in most questions.

Characterizing multi-agent team behavior from partial team tracings : evidence from the English Premier League

Real-world AI systems have been recently deployed which can automatically analyze the plan and tactics of tennis players. As the game-state is updated regularly at short intervals (i.e. point-level), a library of successful and unsuccessful plans of a player can be learnt over time. Given the relative strengths and weaknesses of a player’s plans, a set of proven plans or tactics from the library that characterize a player can be identified. For low-scoring, continuous team sports like soccer, such analysis for multi-agent teams does not exist as the game is not segmented into “discretized” plays (i.e. plans), making it difficult to obtain a library that characterizes a team’s behavior. Additionally, as player tracking data is costly and difficult to obtain, we only have partial team tracings in the form of ball actions which makes this problem even more difficult. In this paper, we propose a method to overcome these issues by representing team behavior via play-segments, which are spatio-temporal descriptions of ball movement over fixed windows of time. Using these representations we can characterize team behavior from entropy maps, which give a measure of predictability of team behaviors across the field. We show the efficacy and applicability of our method on the 2010-2011 English Premier League soccer data.