Enlarging the Margins in Perceptron Decision Trees

Capacity control in perceptron decision trees is typically performed by controlling their size. We prove that other quantities can be as relevant to reduce their flexibility and combat overfitting. In particular, we provide an upper bound on the generalization error which depends both on the size of the tree and on the margin of the decision nodes. So enlarging the margin in perceptron decision trees will reduce the upper bound on generalization error. Based on this analysis, we introduce three new algorithms, which can induce large margin perceptron decision trees. To assess the effect of the large margin bias, OC1 (Journal of Artificial Intelligence Research, 1994, 2, 1–32.) of Murthy, Kasif and Salzberg, a well-known system for inducing perceptron decision trees, is used as the baseline algorithm. An extensive experimental study on real world data showed that all three new algorithms perform better or at least not significantly worse than OC1 on almost every dataset with only one exception. OC1 performed worse than the best margin-based method on every dataset.     

Optimized temporal monitors for SystemC

SystemC is a modeling language built as an extension of C++. Its growing popularity and the increasing complexity of designs have motivated research efforts aimed at the verification of SystemC models using assertion-based verification (ABV), where the designer asserts properties that capture the design intent in a formal language such as PSL or SVA. The model then can be verified against the properties using runtime or formal verification techniques. In this paper we focus on automated generation of runtime monitors from temporal properties. Our focus is on minimizing runtime overhead, rather than monitor size or monitor-generation time. We identify four issues in monitor generation: state minimization, alphabet representation, alphabet minimization, and monitor encoding. We conduct extensive experimentation and identify a combination of settings that offers the best performance in terms of runtime overhead.     

Processing aspects in the low-frequency noise of nMOSFETs on strained-silicon substrates

The impact of different processing factors on the low-frequency (LF) noise of nMOSFETs fabricated in strained-silicon (SSi) substrates will be described. It is shown that the use of an SSi substrate can yield improved LF noise performance compared with standard Czochralski silicon material. This is demonstrated for both full-wafer and selective epitaxial SSi material. The lower 1/f noise points to an improved gate oxide quality, i.e., with a lower interface and bulk defect density, and is correlated with the low-field mobility or transconductance of the transistors. At the same time, it will be demonstrated that there exist defect-related LF noise mechanisms, which generally give rise to excess generation-recombination (GR) noise. Associated with this GR noise, a degradation of either the OFF-state leakage current or the mobility (transconductance) of the devices is observed. It is clear that noise is a sensitive parameter to local defectiveness and may be a useful tool for both materials' characterization and the analysis of processing-related device degradation mechanisms.     

The LabMatrix system

The LabMatrix^™ is a prototyping system designed to give the user a practical and versatile platform for testing microfluidic applications in the fields of health care and life sciences. The LabMatrix^™ system consists of a microfluidic breadboard and cover that align and secure a series of specially designed LabMatrix^™ microfluidic chips. Chips are easily arranged and rearranged into a user-defined fluidic network. The LabMatrix^™ system is designed with maximum flexibility in mind, providing the user with a means to prototype a wide range of microfluidic applications in a short period.     

Counting Fluorescent Dye Molecules on DNA Origami by Means of Photon Statistics

Obtaining quantitative information about molecular assemblies with high spatial and temporal resolution is a challenging task in fluorescence microscopy. Single‐molecule techniques build on the ability to count molecules one by one. Here, a method is presented that extends recent approaches to analyze the statistics of coincidently emitted photons to enable reliable counting of molecules in the range of 1–20. This method does not require photochemistry such as blinking or bleaching. DNA origami structures are labeled with up to 36 dye molecules as a new evaluation tool to characterize this counting by a photon statistics approach. Labeled DNA origami has a well‐defined labeling stoichiometry and ensures equal brightness for all dyes incorporated. Bias and precision of the estimating algorithm are determined, along with the minimal acquisition time required for robust estimation. Complexes containing up to 18 molecules can be investigated non‐invasively within 150 ms. The method might become a quantifying add‐on for confocal microscopes and could be especially powerful in combination with STED/RESOLFT‐type microscopy.     

Not so black and white: Memory for ambiguous group members.

Exponential increases in multiracial identities, expected over the next century, create a conundrum for perceivers accustomed to classifying people as their own- or other-race. The current research examines how perceivers resolve this dilemma with regard to the own-race bias. The authors hypothesized that perceivers are not motivated to include ambiguous-race individuals in the in-group and therefore have some difficulty remembering these individuals. Both racially ambiguous and other-race faces were misremembered more often than own-race faces (Study 1), though memory for ambiguous faces was improved among perceivers motivated to include biracial individuals in the in-group (Study 2). Racial labels assigned to racially ambiguous faces determined memory for these faces, suggesting that uncertainty provides the motivational context for discounting ambiguous faces in memory (Study 3). Finally, an inclusion motivation fostered cognitive associations between racially ambiguous faces and the in-group. Moreover, the extent to which perceivers associated racially ambiguous faces with the in-group predicted memory for ambiguous faces and accounted for the impact of motivation on memory (Study 4). Thus, memory for biracial individuals seems to involve a flexible person construal process shaped by motivational factors. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Underage drinking among young adolescent girls: The role of family processes.

Guided by family interaction theory, this study examined the influences of psychological, peer, and familial processes on alcohol use among young adolescent girls and assessed the contributions of familial factors. An ethnically diverse sample of 1,187 pairs of girls (M age = 12.83 years), and their mothers completed surveys online. Questionnaires assessed girls’ lifetime and recent alcohol use, as well as girls’ demographic, psychological, peer, and family characteristics. Hierarchical logistic regression models showed that although girls’ drinking was associated with a number of psychological and peer factors, the contributions of family domain variables to girls’ drinking were above and beyond that of psychological and peer factors. The interaction analyses further highlighted that having family rules, high family involvement, and greater family communication may offset risks in psychological and peer domains. Study findings underscore the multifaceted etiology of drinking among young adolescent girls and assert the crucial roles of familial processes. Prevention programs should be integrative, target processes at multiple domains, and include work with parents. (PsycINFO Database Record (c) 2010 APA, all rights reserved)