Improved generative semisupervised learning based on finely grained component-conditional class labeling

We introduce new inductive, generative semisupervised mixtures with more finely grained class label generation mechanisms than in previous work. Our models combine advantages of semisupervised mixtures, which achieve label extrapolation over a component, and nearest-neighbor (NN)/nearest-prototype (NP) classification, which achieve accurate classification in the vicinity of labeled samples or prototypes. For our NN-based method, we propose a novel two-stage stochastic data generation, with all samples first generated using a standard finite mixture and then all class labels generated, conditioned on the samples and their components of origin. This mechanism entails an underlying Markov random field, specific to each mixture component or cluster. We invoke the pseudo-likelihood formulation, which forms the basis for an approximate generalized expectation-maximization model learning algorithm. Our NP-based model overcomes a problem with the NN-based model that manifests at very low labeled fractions. Both models are advantageous when within-component class proportions are not constant over the feature space region "owned by" a component. The practicality of this scenario is borne out by experiments on UC Irvine data sets, which demonstrate significant gains in classification accuracy over previous semisupervised mixtures and also overall gains, over KNN classification. Moreover, for very small labeled fractions, our methods overall outperform supervised linear and nonlinear kernel support vector machines.     

Hybrid method for designing digital Butterworth filters

A procedure for designing digital Butterworth filters is proposed. The procedure determines the denominator and the numerator of the filter transfer function based on the positions of the poles in the s-plane and zeros in the z-plane, respectively, and calculates the gain factor using a maximum point normalization method. In contrast to some conventional algorithms, the presented procedure is much simpler by directly obtaining the filter with 3-dB frequencies. This makes the presented algorithm a useful tool for determining the boundaries in electronic or communication systems’ frequency responses. Moreover, the proposed algorithm is compatible with high-order transformations which are the limitations of general pole-zero placement techniques. The proposed method is illustrated by the examples of designing the low-pass, high-pass, band-pass, and band-stop filter.     

The influence of skin depth on crack measurement by the ac field technique

A general solution is given to describe the ac field around surface-breaking cracks for arbitrary values of the ratio of skin depth δ to crack depthd. This solution allows accurate interpretation of crack depth from voltage readings taken by a Crack Microgauge instrument in cases where δ andd are comparable. It is shown to agree with asymptotic formulae obtained for the cases where δ/d is very small and very large. Provided that the probe length Δ is large compared with δ, the calculations show that a very good approximation to the crack depth may be obtained, irrespective of the ratio δ/d, by using the thin-skin asymptotic formula $$d = \frac{1}{2}\Delta (V_2 - V_1 )/V_1 $$ whereV ₁ andV ₂ are probe voltages registered at positions just off and across the crack, respectively. The problem is solved for a crack of infinite aspect ratio which has no opening, but it can be applied with confidence at the centerline of cracks of fairly large aspect ratio.     

The Morning After: Limit Violations and the Self-Regulation of Alcohol Consumption.

The authors examined whether the reciprocal relationship between alcohol consumption and distress unfolded over time in 2 samples of social drinkers. Participants monitored their alcohol intake and their cognitive and emotional responses to that drinking on hand-held computers. On mornings after drinking, those who had violated their self-imposed limits the day before reported more guilt, even after controlling for acute negative symptoms of drinking and amount consumed. Reciprocally, guilt led to poorer self-regulation of alcohol intake: Greater distress over alcohol consumption was linked to more intake, intoxication, and more limit violations. Individual differences moderated the relationships among limit violations, distress, and drinking. Consistent with the limit violation effect, violating a limit produced distress over consumption among social drinkers, and they responded to that distress by drinking more. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Fuzzy PI control design for an industrial weigh belt feeder

An industrial weigh belt feeder is used to transport solid materials into a manufacturing process at a constant feedrate. It exhibits nonlinear behavior because of motor friction, saturation, and quantization noise in the sensors, which makes standard autotuning methods difficult to implement. The paper proposes and experimentally demonstrates two types of fuzzy logic controllers for an industrial weigh belt feeder. The first type is a PI-like fuzzy logic controller (FLC). A gain scheduled PI-like FLC and a self-tuning PI-like FLC are presented. For the gain scheduled PI-like FLC the output scaling factor of the controller is gain scheduled with the change of setpoint. For the self-tuning PI-like FLC, the output scaling factor of the controller is modified online by an updating factor whose value is determined by a rule base with the error and change of error of the controlled variable as the inputs. A fuzzy PI controller is also presented, where the proportional and integral gains are tuned online based on fuzzy inference rules. Experimental results show the effectiveness of the proposed fuzzy logic controllers. A performance comparison of the three controllers is also given.     

The Challenges of Cross-Modal Translation: English-to-Sign-Language Translation in the Zardoz System

The sign languages used by deaf communities around the world represent a linguistic challenge that natural-language researchers in AI have only recently begun to take up. This challenge is particularly relevant to research in Machine Translation (MT), as natural sign languages have evolved in deaf communities into efficient modes of gestural communication, which differ from English not only in modality but in grammatical structure, exploiting a higher dimensionality of spatial expression. In this paper we describe Zardoz, an on-going AI research system that tackles the cross-modal MT problem, translating English text into fluid sign language. The paper presents an architectural overview of Zardoz, describing its central blackboard organization, the nature of its interlingual representation, and the major components which interact through this blackboard both to analyze the verbal input and generate the corresponding gestural output in one of a number of sign variants.     

Single-factor coefficient bounds

In a 1993 paper Beauzamy, Trevisan and Wang derived a single-factor coefficient bound, one which limits the max norm (height) of at least one irreducible factor of any univariate integral polynomial A. Their bound is a function of the degree and the weighted norm of A. In the conclusion of their paper they ask whether the max norm of A might already be a single-factor coefficient bound. In 1998 Knuth, citing these authors, asked instead whether there is a constant c such that c times the max norm of A is a single-factor coefficient bound. We present the results of extensive calculations relating to this question. We show that c, if it exists, must be greater than 2 and accrue evidence in support of a conjecture that the answer to Knuth’s question is “no”.     

In situ nitridation of titanium–molybdenum alloys during laser deposition

In situ nitridation during laser deposition of titanium–molybdenum alloys from elemental powder blends has been achieved by introducing the reactive nitrogen gas during the deposition process. Thus, Ti–Mo–N alloys have been deposited using the laser engineered net shaping (LENS^TM) process and resulted in the formation of a hard α(Ti,N) phase, exhibiting a dendritic morphology, distributed within a β(Ti–Mo) matrix with fine scale transformed α precipitates. Varying the composition of the Ar + N₂ gas employed during laser deposition permits a systematic increase in the nitrogen content of the as-deposited Ti–Mo–N alloy. Interestingly, the addition of nitrogen, which stabilizes the α phase in Ti, changes the solidification pathway and the consequent sequence of phase evolution in these alloys. The nitrogen-enriched hcp α(Ti,N) phase has higher c/a ratio, exhibits an equiaxed morphology, and tends to form in clusters separated by ribs of the Mo-rich β phase. The Ti–Mo–N alloys also exhibit a substantial enhancement in microhardness due to the formation of this α(Ti,N) phase, combining it with the desirable properties of the β-Ti matrix, such as excellent ductility, toughness, and formability.     

Do problematic and non-problematic video game players differ in extraversion,trait empathy,social capital and prosocial tendencies?

The increasing popularity of online and multiplayer games has meant that for many, social interaction and cooperation are vital components of the gaming experience. Previous research has suggested that not only has this made gaming more attractive to socially oriented people but also that it may be socially beneficial in terms of social capital and prosocial behaviors. However, for problematic video game players (those showing signs of compulsive or detrimental video game use), this may not be the case, and a number of theories hold deficiencies in socializing in real life as central to the development of this issue. In the present study, an online questionnaire completed by 416 participants assessed problematic video game use, extraversion, trait empathy, online and offline social capital and prosocial tendencies. Contrary to hypotheses, non-problematic, problematic and non-gamers did not differ in empathy, extraversion or prosocial tendencies. Problematic video game play was, however, associated with significantly higher online social capital and lower offline social capital whereas non-problematic players demonstrated only higher online capital than non-players. This highlights the importance of social support but suggests personality is not an influential factor.     

Stability of discrete-time bilinear systems with constant inputs

In this paper discrete-time bilinear systems are considered. Necessary and sufficient conditions are developed which ensure that for any constant input in an unbounded control domain, the trajectories of the system tend to the origin as time increases.