A Novel Parameter Estimation Method for Boltzmann Machines.

We propose a novel estimator for a specific class of probabilistic models on discrete spaces such as the Boltzmann machine. The proposed estimator is derived from minimization of a convex risk function and can be constructed without calculating the normalization constant, whose computational cost is exponential order. We investigate statistical properties of the proposed estimator such as consistency and asymptotic normality in the framework of the estimating function. Small experiments show that the proposed estimator can attain comparable performance to the maximum likelihood expectation at a much lower computational cost and is applicable to high-dimensional data.

Causal Relationships between the Psychological Acceptance Process of Athletic Injury and Athletic Rehabilitation Behavior.

[Purpose] The purpose of this study was to examine the causal relationships between the psychological acceptance process of athletic injury and athletic-rehabilitation behavior. [Subjects] One hundred forty-four athletes who had injury experiences participated in this study, and 133 (mean age = 20.21 years, SD = 1.07; mean weeks without playing sports = 7.97 weeks, SD = 11.26) of them provided valid questionnaire responses which were subjected to analysis. [Methods] The subjects were asked to answer our originally designed questionnaire, the Psychosocial Recovery Factor Scale (PSRF-S), and two other pre-existing scales, the Athletic Injury Psychological Acceptance Scale and the Athletic-Rehabilitation Dedication Scale. [Results] The results of factor analysis indicate "emotional stability", "social competence in the team", "temporal perspective", and "communication with the teammates" are factors of the PSRF-S. Lastly, the causal model in which psychosocial recovery factors are mediated by psychological acceptance of athletic injury, and influence on rehabilitation behaviors, was examined using structural equation modeling (SEM). The results of SEM indicate that the factors of emotional stability and temporal perspective are mediated by the psychological acceptance of the injury, which positively influences athletic-rehabilitation dedication. [Conclusion] The causal model was confirmed to be valid.

An extension of the receiver operating characteristic curve and AUC-optimal classification.

While most proposed methods for solving classification problems focus on minimization of the classification error rate, we are interested in the receiver operating characteristic (ROC) curve, which provides more information about classification performance than the error rate does. The area under the ROC curve (AUC) is a natural measure for overall assessment of a classifier based on the ROC curve. We discuss a class of concave functions for AUC maximization in which a boosting-type algorithm including RankBoost is considered, and the Bayesian risk consistency and the lower bound of the optimum function are discussed. A procedure derived by maximizing a specific optimum function has high robustness, based on gross error sensitivity. Additionally, we focus on the partial AUC, which is the partial area under the ROC curve. For example, in medical screening, a high true-positive rate to the fixed lower false-positive rate is preferable and thus the partial AUC corresponding to lower false-positive rates is much more important than the remaining AUC. We extend the class of concave optimum functions for partial AUC optimality with the boosting algorithm. We investigated the validity of the proposed method through several experiments with data sets in the UCI repository.

Graph-based composite local Bregman divergences on discrete sample spaces.

This paper develops a general framework of statistical inference on discrete sample spaces, on which a neighborhood system is defined by an undirected graph. The scoring rule is a measure of the goodness of fit for the model to observed samples, and we employ its localized version, local scoring rules, which does not require the normalization constant. We show that the local scoring rule is closely related to a discrepancy measure called composite local Bregman divergence. Then, we investigate the statistical consistency of local scoring rules in terms of the graphical structure of the sample space. Moreover, we propose a robust and computationally efficient estimator based on our framework. In numerical experiments, we investigate the relation between the neighborhood system and estimation accuracy. Also, we numerically evaluate the robustness of localized estimators.

Relationship of sports experience and ego development of adolescent Japanese athletes.

This study examined the relationship of sports experience with ego development. A questionnaire was used to assess experience of Crisis, Exploration, and Commitment in the issues of Athletic Performance and of Being a Teammate in 782 adolescent Japanese athletes (423 boys, M age = 15.2 yr.; 359 girls, M age = 15.0 yr.). Their Ego Levels were assessed using the Washington University Sentence Completion Test. Correlations indicated that scores on Crisis, Exploration, and Commitment in the issues of Athletic Performance and Being a Teammate were generally associated with Ego Development. Multiple regression analyses indicated that, for boys, the issue of Athletic Performance was closely associated with Ego Development, while for girls, the issue of Being a Teammate was closely associated with Ego Development. Sports experience with crisis, exploration, and commitment may be related to accommodation, which is, in turn, related to ego development. The sex differences on issues related to ego development may be associated with differences in sex-role development for boys and girls.

A multiclass classification method based on decoding of binary classifiers.

In this letter, we present new methods of multiclass classification that combine multiple binary classifiers. Misclassification of each binary classifier is formulated as a bit inversion error with probabilistic models by making an analogy to the context of information transmission theory. Dependence between binary classifiers is incorporated into our model, which makes a decoder a type of Boltzmann machine. We performed experimental studies using a synthetic data set, data sets from the UCI repository, and bioinformatics data sets, and the results show that the proposed methods are superior to the existing multiclass classification methods.

Robust model selection for classification of microarrays.

Recently, microarray-based cancer diagnosis systems have been increasingly investigated. However, cost reduction and reliability assurance of such diagnosis systems are still remaining problems in real clinical scenes. To reduce the cost, we need a supervised classifier involving the smallest number of genes, as long as the classifier is sufficiently reliable. To achieve a reliable classifier, we should assess candidate classifiers and select the best one. In the selection process of the best classifier, however, the assessment criterion must involve large variance because of limited number of samples and non-negligible observation noise. Therefore, even if a classifier with a very small number of genes exhibited the smallest leave-one-out cross-validation (LOO) error rate, it would not necessarily be reliable because classifiers based on a small number of genes tend to show large variance. We propose a robust model selection criterion, the min-max criterion, based on a resampling bootstrap simulation to assess the variance of estimation of classification error rates. We applied our assessment framework to four published real gene expression datasets and one synthetic dataset. We found that a state-of-the-art procedure, weighted voting classifiers with LOO criterion, had a non-negligible risk of selecting extremely poor classifiers and, on the other hand, that the new min-max criterion could eliminate that risk. These finding suggests that our criterion presents a safer procedure to design a practical cancer diagnosis system.

Robust boosting algorithm against mislabeling in multiclass problems.

We discuss robustness against mislabeling in multiclass labels for classification problems and propose two algorithms of boosting, the normalized Eta-Boost.M and Eta-Boost.M, based on the Eta-divergence. Those two boosting algorithms are closely related to models of mislabeling in which the label is erroneously exchanged for others. For the two boosting algorithms, theoretical aspects supporting the robustness for mislabeling are explored. We apply the proposed two boosting methods for synthetic and real data sets to investigate the performance of these methods, focusing on robustness, and confirm the validity of the proposed methods.

Robust loss functions for boosting.

Boosting is known as a gradient descent algorithm over loss functions. It is often pointed out that the typical boosting algorithm, Adaboost, is highly affected by outliers. In this letter, loss functions for robust boosting are studied. Based on the concept of robust statistics, we propose a transformation of loss functions that makes boosting algorithms robust against extreme outliers. Next, the truncation of loss functions is applied to contamination models that describe the occurrence of mislabels near decision boundaries. Numerical experiments illustrate that the proposed loss functions derived from the contamination models are useful for handling highly noisy data in comparison with other loss functions.

In vitro and in vivo antibacterial activities of CS-023 (RO4908463), a novel parenteral carbapenem.

CS-023 (RO4908463, formerly R-115685) is a novel 1beta-methylcarbapenem with 5-substituted pyrrolidin-3-ylthio groups, including an amidine moiety at the C-2 position. Its antibacterial activity was tested against 1,214 clinical isolates of 32 species and was compared with those of imipenem, meropenem, ceftazidime, ceftriaxone, ampicillin, amikacin, and levofloxacin. CS-023 exhibited a broad spectrum of activity against gram-positive and -negative aerobes and anaerobes, including methicillin-resistant Staphylococcus aureus (MRSA), methicillin-resistant Staphylococcus epidermidis, penicillin-resistant Streptococcus pneumoniae (PRSP), beta-lactamase-negative ampicillin-resistant Haemophilus influenzae, and Pseudomonas aeruginosa. CS-023 showed the most potent activity among the compounds tested against P. aeruginosa and MRSA, with MICs at which 90% of isolates tested were inhibited of 4 microg/ml and 8 microg/ml, respectively. CS-023 was stable against hydrolysis by the beta-lactamases from Enterobacter cloacae and Proteus vulgaris. CS-023 also showed potent activity against extended-spectrum beta-lactamase-producing Escherichia coli. The in vivo efficacy of CS-023 was evaluated with a murine systemic infection model induced by 13 strains of gram-positive and -negative pathogens and a lung infection model induced by 2 strains of PRSP (serotypes 6 and 19). Against the systemic infections with PRSP, MRSA, and P. aeruginosa and the lung infections, the efficacy of CS-023 was comparable to those of imipenem/cilastatin and vancomycin (tested against lung infections only) and superior to those of meropenem, ceftriaxone, and ceftazidime (tested against P. aeruginosa infections only). These results suggest that CS-023 has potential for the treatment of nosocomial bacterial infections by gram-positive and -negative pathogens, including MRSA and P. aeruginosa.

Robustifying AdaBoost by adding the naive error rate.

AdaBoost can be derived by sequential minimization of the exponential loss function. It implements the learning process by exponentially reweighting examples according to classification results. However, weights are often too sharply tuned, so that AdaBoost suffers from the nonrobustness and overlearning. Wepropose a new boosting method that is a slight modification of AdaBoost. The loss function is defined by a mixture of the exponential loss and naive error loss functions. As a result, the proposed method incorporates the effect of forgetfulness into AdaBoost. The statistical significance of our method is discussed, and simulations are presented for confirmation.

Properties of extended-spectrum beta-lactamases constructed by site-directed mutagenesis.

Plasmids carrying three types of TEM-type extended-spectrum beta-lactamase (ESBL) genes, encoding TEM-3, TEM-5, and TEM-9, respectively, were constructed by site-directed mutagenesis. ESBL producers were prepared by transformation of Escherichia coli JM109 with a plasmid carrying one gene of either the three TEM types, an SHV-type, or a Toho-1 group gene. This strategy with the same vector and host strain can exclude the contribution of other factors to susceptibility, and is useful in Japan, where few TEM-type ESBL producers have been isolated. In vitro antibacterial activities of 23 beta-lactam antibiotics were tested against the ESBL producers by the agar dilution method, and the results were compared. The minimum inhibitory concentrations (MICs) of penicillins tested were more than 32 micro g/ml against both the parental RTEM and ESBL producers, but they were substantially decreased by a combination with beta-lactamase inhibitors. Compared with the MICs against the ESBL-nonproducing host strain, the MICs of the cephalosporins tested for the ESBL producers were increased more than eight times in most cases and in several cases soared to more than 2048 times against a Toho-1 ESBL producer. On the other hand, the MICs of carbapenem, cephamycin, and penem antibiotics were generally comparable to those against the host strain, and were increased by 32 times at most. Kinetic analysis revealed that extended-spectrum cephalosporins were hydrolyzed only slightly to moderately by the TEM-type ESBLs, while carbapenems and a cephamycin were scarcely hydrolyzed, and rather inhibited or inactivated the mutant enzymes.

Information geometry of U-Boost and Bregman divergence.

We aim at an extension of AdaBoost to U-Boost, in the paradigm to build a stronger classification machine from a set of weak learning machines. A geometric understanding of the Bregman divergence defined by a generic convex function U leads to the U-Boost method in the framework of information geometry extended to the space of the finite measures over a label set. We propose two versions of U-Boost learning algorithms by taking account of whether the domain is restricted to the space of probability functions. In the sequential step, we observe that the two adjacent and the initial classifiers are associated with a right triangle in the scale via the Bregman divergence, called the Pythagorean relation. This leads to a mild convergence property of the U-Boost algorithm as seen in the expectation-maximization algorithm. Statistical discussions for consistency and robustness elucidate the properties of the U-Boost methods based on a stochastic assumption for training data.