The exact sample complexity of PAC-learning problems with unit VC dimension

The Vapnik-Chervonenkis (VC) dimension is a combinatorial measure of a certain class of machine learning problems, which may be used to obtain upper and lower bounds on the number of training examples needed to learn to prescribed levels of accuracy. Most of the known bounds apply to the Probably Approximately Correct (PAC) framework, which is the framework within which we work in this paper. For a learning problem with some known VC dimension, much is known about the order of growth of the sample-size requirement of the problem, as a function of the PAC parameters. The exact value of sample-size requirement is however less well-known, and depends heavily on the particular learning algorithm being used. This is a major obstacle to the practical application of the VC dimension. Hence it is important to know exactly how the sample-size requirement depends on VC dimension, and with that in mind, we describe a general algorithm for learning problems having VC dimension 1. Its sample-size requirement is minimal (as a function of the PAC parameters), and turns out to be the same for all non-trivial learning problems having VC dimension 1. While the method used cannot be naively generalised to higher VC dimension, it suggests that optimal algorithm-dependent bounds may improve substantially on current upper bounds.

Reinforcement Learning by Policy Search

One objective of artificial intelligence is to model the behavior of an intelligent agent interacting with its environment. The environment's transformations can be modeled as a Markov chain, whose state is partially observable to the agent and affected by its actions; such processes are known as partially observable Markov decision processes (POMDPs). While the environment's dynamics are assumed to obey certain rules, the agent does not know them and must learn. In this dissertation we focus on the agent's adaptation as captured by the reinforcement learning framework. This means learning a policy---a mapping of observations into actions---based on feedback from the environment. The learning can be viewed as browsing a set of policies while evaluating them by trial through interaction with the environment. The set of policies is constrained by the architecture of the agent's controller. POMDPs require a controller to have a memory. We investigate controllers with memory, including controllers with external memory, finite state controllers and distributed controllers for multi-agent systems. For these various controllers we work out the details of the algorithms which learn by ascending the gradient of expected cumulative reinforcement. Building on statistical learning theory and experiment design theory, a policy evaluation algorithm is developed for the case of experience re-use. We address the question of sufficient experience for uniform convergence of policy evaluation and obtain sample complexity bounds for various estimators. Finally, we demonstrate the performance of the proposed algorithms on several domains, the most complex of which is simulated adaptive packet routing in a telecommunication network.

Sequential Optimal Recovery: A Paradigm for Active Learning

In most classical frameworks for learning from examples, it is assumed that examples are randomly drawn and presented to the learner. In this paper, we consider the possibility of a more active learner who is allowed to choose his/her own examples. Our investigations are carried out in a function approximation setting. In particular, using arguments from optimal recovery (Micchelli and Rivlin, 1976), we develop an adaptive sampling strategy (equivalent to adaptive approximation) for arbitrary approximation schemes. We provide a general formulation of the problem and show how it can be regarded as sequential optimal recovery. We demonstrate the application of this general formulation to two special cases of functions on the real line 1) monotonically increasing functions and 2) functions with bounded derivative. An extensive investigation of the sample complexity of approximating these functions is conducted yielding both theoretical and empirical results on test functions. Our theoretical results (stated insPAC-style), along with the simulations demonstrate the superiority of our active scheme over both passive learning as well as classical optimal recovery. The analysis of active function approximation is conducted in a worst-case setting, in contrast with other Bayesian paradigms obtained from optimal design (Mackay, 1992).

A Formulation for Active Learning with Applications to Object Detection

We discuss a formulation for active example selection for function learning problems. This formulation is obtained by adapting Fedorov's optimal experiment design to the learning problem. We specifically show how to analytically derive example selection algorithms for certain well defined function classes. We then explore the behavior and sample complexity of such active learning algorithms. Finally, we view object detection as a special case of function learning and show how our formulation reduces to a useful heuristic to choose examples to reduce the generalization error.

A functional proteomic method for the enrichment of peripheral membrane proteins reveals the collagen binding protein Hsp47 is exposed on the surface of activated human platelets

Platelets are small blood cells vital for hemostasis. Following vascular damage, platelets adhere to collagens and activate, forming a thrombus that plugs the wound and prevents blood loss. Stimulation of the platelet collagen receptor glycoprotein VI (GPVI) allows recruitment of proteins to receptor-proximal signaling complexes on the inner-leaflet of the plasma membrane. These proteins are often present at low concentrations; therefore, signaling-complex characterization using mass spectrometry is limited due to high sample complexity. We describe a method that facilitates detection of signaling proteins concentrated on membranes. Peripheral membrane proteins (reversibly associated with membranes) were eluted from human platelets with alkaline sodium carbonate. Liquid-phase isoelectric focusing and gel electrophoresis were used to identify proteins that changed in levels on membranes from GPVI-stimulated platelets. Immunoblot analysis verified protein recruitment to platelet membranes and subsequent protein phosphorylation was preserved. Hsp47, a collagen binding protein, was among the proteins identified and found to be exposed on the surface of GPVI-activated platelets. Inhibition of Hsp47 abolished platelet aggregation in response to collagen, while only partially reducing aggregation in response to other platelet agonists. We propose that Hsp47 may therefore play a role in hemostasis and thrombosis.

Complexity quantification of dense array EEG using sample entropy analysis

n this paper, a time series complexity analysis of dense array electroencephalogram signals is carried out using the recently introduced Sample Entropy (SampEn) measure. This statistic quantifies the regularity in signals recorded from systems that can vary from the purely deterministic to purely stochastic realm. The present analysis is conducted with an objective of gaining insight into complexity variations related to changing brain dynamics for EEG recorded from the three cases of passive, eyes closed condition, a mental arithmetic task and the same mental task carried out after a physical exertion task. It is observed that the statistic is a robust quantifier of complexity suited for short physiological signals such as the EEG and it points to the specific brain regions that exhibit lowered complexity during the mental task state as compared to a passive, relaxed state. In the case of mental tasks carried out before and after the performance of a physical exercise, the statistic can detect the variations brought in by the intermediate fatigue inducing exercise period. This enhances its utility in detecting subtle changes in the brain state that can find wider scope for applications in EEG based brain studies.

Evaluation of physiologic complexity in time series using generalized sample entropy and surrogate data analysis

Complexity in time series is an intriguing feature of living dynamical systems, with potential use for identification of system state. Although various methods have been proposed for measuring physiologic complexity, uncorrelated time series are often assigned high values of complexity, errouneously classifying them as a complex physiological signals. Here, we propose and discuss a method for complex system analysis based on generalized statistical formalism and surrogate time series. Sample entropy (SampEn) was rewritten inspired in Tsallis generalized entropy, as function of q parameter (qSampEn). qSDiff curves were calculated, which consist of differences between original and surrogate series qSampEn. We evaluated qSDiff for 125 real heart rate variability (HRV) dynamics, divided into groups of 70 healthy, 44 congestive heart failure (CHF), and 11 atrial fibrillation (AF) subjects, and for simulated series of stochastic and chaotic process. The evaluations showed that, for nonperiodic signals, qSDiff curves have a maximum point (qSDiff(max)) for q not equal 1. Values of q where the maximum point occurs and where qSDiff is zero were also evaluated. Only qSDiff(max) values were capable of distinguish HRV groups (p-values 5.10 x 10(-3); 1.11 x 10(-7), and 5.50 x 10(-7) for healthy vs. CHF, healthy vs. AF, and CHF vs. AF, respectively), consistently with the concept of physiologic complexity, and suggests a potential use for chaotic system analysis. (C) 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4758815]

Disordered eating, weight dissatisfaction and dieting in relation to unwanted childhood sexual experiences in a community sample

A study was conducted to examine the relationships among eating pathology, weight dissatisfaction and dieting, and unwanted sexual experiences in childhood. An unselected community sample of 201 young and 268 middle-aged women were administered questionnaires assessing eating behaviors and attitudes, and past and current sexual abuse. Results showed differential relationships among these factors for the two age cohorts: for young women, past sexual abuse predicted weight dissatisfaction, but not dieting or disordered eating behaviors, whereas for middle-aged women, past abuse was predictive of disordered eating, but not dieting or weight dissatisfaction. Current physical or sexual abuse was also found to be predictive of disordered eating for the young women. These findings underscore the complexity of the relationships among unwanted sexual experiences and eating and weight pathology, and suggest that the timing of sexual abuse, and the age of the woman, are important mediating factors. (C) 1998 Elsevier Science Inc.

ADHD in a representative sample of the Brazilian population: estimated prevalence and comparative adequacy of criteria between adolescents and adults according to the item response theory

Background: Attention deficit hyperactivity disorder (ADHD) is a clinically significant disorder in adulthood, but current diagnostic criteria and instruments do not seem to adequately capture the complexity of the disorder in this developmental phase. Accordingly, there are limited data on the proportion of adults affected by the disorder, specially in developing countries. Method: We assessed a representative household sample of the Brazilian population for ADHD with the Adult ADHD Self-report Scale (ASRS) Screener, and evaluated the instrument according to the Rasch model of item response theory. Results: The sample was comprised by 3007 individuals, and the overal prevalence of positive screeners for ADHD was 5.8% [95% confidence interval (CI), 4.8-7.0]. Rasch analyses revealed the misfitt of the overall sample to expectations of the model. The evaluation of the sample stratified by age revealed that data for adolescents showed a signficant fittnes to the model expectations, while items completed by adults were not adequated. Conclusions: The lack of fitness to the model for adult respondents challenges the possibility of a linear transformation of the ordinal data into interval measures and the utilization of parametric analyses of data. This result suggests that diagnostic criteria and instruments for adult ADHD must take into account a developmental perspective. Moreover, it calls for further evaluation of currently employed research methods in light of modern theories of psychometrics. Copyright (C) 2010 John Wiley & Sons, Ltd.

Medication regimen complexity in institutionalized elderly people in an aging society

Background: Complex medication regimens may adversely affect compliance and treatment outcomes. Complexity can be assessed with the medication regimen complexity index (MRCI), which has proved to be a valid, reliable tool, with potential uses in both practice and research. Objective: To use the MRCI to assess medication regimen complexity in institutionalized elderly people. Setting: Five nursing homes in mainland Portugal. Methods: A descriptive, cross-sectional study of institutionalized elderly people (n = 415) was performed from March to June 2009, including all inpatients aged 65 and over taking at least one medication per day. Main outcome measure: Medication regimen complexity index. Results: The mean age of the sample was 83.9 years (±6.6 years), and 60.2 % were women. The elderly patients were taking a large number of drugs, with 76.6 % taking more than five medications per day. The average medication regimen complexity was 18.2 (±SD = 9.6), and was higher in the females (p < 0.001). The most decisive factors contributing to the complexity were the number of drugs and dosage frequency. In regimens with the same number of medications, schedule was the most relevant factor in the final score (r = 0.922), followed by pharmaceutical forms (r = 0.768) and additional instructions (r = 0.742). Conclusion: Medication regimen complexity proved to be high. There is certainly potential for the pharmacist's intervention to reduce it as part as the medication review routine in all the patients.

Systematic survey of clonal complexity in tuberculosis at a populational level and detailed characterization of the isolates involved.

Clonally complex infections by Mycobacterium tuberculosis are progressively more accepted. Studies of their dimension in epidemiological scenarios where the infective pressure is not high are scarce. Our study systematically searched for clonally complex infections (mixed infections by more than one strain and simultaneous presence of clonal variants) by applying mycobacterial interspersed repetitive-unit (MIRU)-variable-number tandem-repeat (VNTR) analysis to M. tuberculosis isolates from two population-based samples of respiratory (703 cases) and respiratory-extrapulmonary (R+E) tuberculosis (TB) cases (71 cases) in a context of moderate TB incidence. Clonally complex infections were found in 11 (1.6%) of the respiratory TB cases and in 10 (14.1%) of those with R+E TB. Among the 21 cases with clonally complex TB, 9 were infected by 2 independent strains and the remaining 12 showed the simultaneous presence of 2 to 3 clonal variants. For the 10 R+E TB cases with clonally complex infections, compartmentalization (different compositions of strains/clonal variants in independent infected sites) was found in 9 of them. All the strains/clonal variants were also genotyped by IS6110-based restriction fragment length polymorphism analysis, which split two MIRU-defined clonal variants, although in general, it showed a lower discriminatory power to identify the clonal heterogeneity revealed by MIRU-VNTR analysis. The comparative analysis of IS6110 insertion sites between coinfecting clonal variants showed differences in the genes coding for a cutinase, a PPE family protein, and two conserved hypothetical proteins. Diagnostic delay, existence of previous TB, risk for overexposure, and clustered/orphan status of the involved strains were analyzed to propose possible explanations for the cases with clonally complex infections. Our study characterizes in detail all the clonally complex infections by M. tuberculosis found in a systematic survey and contributes to the characterization that these phenomena can be found to an extent higher than expected, even in an unselected population-based sample lacking high infective pressure.

Evaluation of the Adolescent Drug Abuse Diagnosis instrument in a Swiss sample of drug abusers.

OBJECTIVE: The aim of this study was to evaluate a French language version of the Adolescent Drug Abuse Diagnosis (ADAD) instrument in a Swiss sample of adolescent illicit drug and/or alcohol users. PARTICIPANTS AND SETTING: The participants in the study were 102 French-speaking adolescents aged 13-19 years who fitted the criteria of illicit drug or alcohol use (at least one substance--except tobacco--once a week during the last 3 months). They were recruited in hospitals, institutions and leisure places. Procedure. The ADAD was administered individually by trained psychologists. It was integrated into a broader protocol including alcohol and drug abuse DSM-IV diagnoses, the BDI-13 (Beck Depression Inventory), life events and treatment trajectories. RESULTS: The ADAD appears to show good inter-rater reliability; the subscales showed good internal coherence and the correlations between the composite scores and the severity ratings were moderate to high. Finally, the results confirmed good concurrent validity for three out of eight ADAD dimensions. CONCLUSIONS: The French language version of the ADAD appears to be an adequate instrument for assessing drug use and associated problems in adolescents. Despite its complexity, the instrument has acceptable validity, reliability and usefulness criteria, enabling international and transcultural comparisons.

Estimation of reference curves for the urinary steroid metabolome in the first year of life in healthy children : Tracing the complexity of human postnatal steroidogenesis.

CONTEXT: Complex steroid disorders such as P450 oxidoreductase deficiency or apparent cortisone reductase deficiency may be recognized by steroid profiling using chromatographic mass spectrometric methods. These methods are highly specific and sensitive, and provide a complete spectrum of steroid metabolites in a single measurement of one sample which makes them superior to immunoassays. The steroid metabolome during the fetal-neonatal transition is characterized by (a) the metabolites of the fetal-placental unit at birth, (b) the fetal adrenal androgens until its involution 3-6 months postnatally, and (c) the steroid metabolites produced by the developing endocrine organs. All these developmental events change the steroid metabolome in an age- and sex-dependent manner during the first year of life. OBJECTIVE: The aim of this study was to provide normative values for the urinary steroid metabolome of healthy newborns at short time intervals in the first year of life. METHODS: We conducted a prospective, longitudinal study to measure 67 urinary steroid metabolites in 21 male and 22 female term healthy newborn infants at 13 time-points from week 1 to week 49 of life. Urine samples were collected from newborn infants before discharge from hospital and from healthy infants at home. Steroid metabolites were measured by gas chromatography-mass spectrometry (GC-MS) and steroid concentrations corrected for urinary creatinine excretion were calculated. RESULTS: 61 steroids showed age and 15 steroids sex specificity. Highest urinary steroid concentrations were found in both sexes for progesterone derivatives, in particular 20α-DH-5α-DH-progesterone, and for highly polar 6α-hydroxylated glucocorticoids. The steroids peaked at week 3 and decreased by ∼80% at week 25 in both sexes. The decline of progestins, androgens and estrogens was more pronounced than of glucocorticoids whereas the excretion of corticosterone and its metabolites and of mineralocorticoids remained constant during the first year of life. CONCLUSION: The urinary steroid profile changes dramatically during the first year of life and correlates with the physiologic developmental changes during the fetal-neonatal transition. Thus detailed normative data during this time period permit the use of steroid profiling as a powerful diagnostic tool.

Dynamic patterns of flow in the workplace: characterizing within-individual variability using a complexity science approach

As a result of the growing interest in studying employee well-being as a complex process that portrays high levels of within-individual variability and evolves over time, this present study considers the experience of flow in the workplace from a nonlinear dynamical systems approach. Our goal is to offer new ways to move the study of employee well-being beyond linear approaches. With nonlinear dynamical systems theory as the backdrop, we conducted a longitudinal study using the experience sampling method and qualitative semi-structured interviews for data collection; 6981 registers of data were collected from a sample of 60 employees. The obtained time series were analyzed using various techniques derived from the nonlinear dynamical systems theory (i.e., recurrence analysis and surrogate data) and multiple correspondence analyses. The results revealed the following: 1) flow in the workplace presents a high degree of within-individual variability; this variability is characterized as chaotic for most of the cases (75%); 2) high levels of flow are associated with chaos; and 3) different dimensions of the flow experience (e.g., merging of action and awareness) as well as individual (e.g., age) and job characteristics (e.g., job tenure) are associated with the emergence of different dynamic patterns (chaotic, linear and random).

Multimorbidity patterns in a national representative sample of the Spanish adult population.

BACKGROUND: In the context of population aging, multimorbidity has emerged as a growing concern in public health. However, little is known about multimorbidity patterns and other issues surrounding chronic diseases. The aim of our study was to examine multimorbidity patterns, the relationship between physical and mental conditions and the distribution of multimorbidity in the Spanish adult population. METHODS: Data from this cross-sectional study was collected from the COURAGE study. A total of 4,583 participants from Spain were included, 3,625 aged over 50. An exploratory factor analysis was conducted to detect multimorbidity patterns in the population over 50 years of age. Crude and adjusted binary logistic regressions were performed to identify individual associations between physical and mental conditions. RESULTS: THREE MULTIMORBIDITY PATTERNS ROSE: 'cardio-respiratory' (angina, asthma, chronic lung disease), 'mental-arthritis' (arthritis, depression, anxiety) and the 'aggregated pattern' (angina, hypertension, stroke, diabetes, cataracts, edentulism, arthritis). After adjusting for covariates, asthma, chronic lung disease, arthritis and the number of physical conditions were associated with depression. Angina and the number of physical conditions were associated with a higher risk of anxiety. With regard to multimorbidity distribution, women over 65 years suffered from the highest rate of multimorbidity (67.3%). CONCLUSION: Multimorbidity prevalence occurs in a high percentage of the Spanish population, especially in the elderly. There are specific multimorbidity patterns and individual associations between physical and mental conditions, which bring new insights into the complexity of chronic patients. There is need to implement patient-centered care which involves these interactions rather than merely paying attention to individual diseases.