A PAC-Bayesian margin bound for linear classifiers

We present a bound on the generalization error of linear classifiers in terms of a refined margin quantity on the training sample. The result is obtained in a probably approximately correct (PAC)-Bayesian framework and is based on geometrical arguments in the space of linear classifiers. The new bound constitutes an exponential improvement of the so far tightest margin bound, which was developed in the luckiness framework, and scales logarithmically in the inverse margin. Even in the case of less training examples than input dimensions sufficiently large margins lead to nontrivial bound values and-for maximum margins-to a vanishing complexity term. In contrast to previous results, however, the new bound does depend on the dimensionality of feature space. The analysis shows that the classical margin is too coarse a measure for the essential quantity that controls the generalization error: the fraction of hypothesis space consistent with the training sample. The practical relevance of the result lies in the fact that the well-known support vector machine is optimal with respect to the new bound only if the feature vectors in the training sample are all of the same length. As a consequence, we recommend to use support vector machines (SVMs) on normalized feature vectors only. Numerical simulations support this recommendation and demonstrate that the new error bound can be used for the purpose of model selection.     

Cryptanalysis of Polly Cracker

An attack on the public key cryptosystem Polly Cracker is described, that reveals the complete secret key σ ∈ F_q ⁿ by means of n (nonadaptively) chosen "fake" ciphertexts     

Retrievals of column water vapor using millimeter-wave radiometric measurements

The airborne Millimeter-wave Imaging Radiometer (MIR) measurements conducted over the Midwest region of the continental United States during January/February 1997 and over the Alaska-Arctic region during May 1998 are used to estimate column water vapor W<0.8 g/cm/sup 2/ under a clear sky. On board the same aircraft are two other instruments, the Cloud Lidar System (CLS) and MODerate-resolution Imaging Spectrometer (MODIS) Airborne Simulator (MAS), which provide cloud cover information and independent measurements of W, respectively. The MIR-estimated W values are compared and found to be in very good agreement with those measured by rawinsondes at near concurrence. A close correlation is found between the MIR-estimated W and that estimated from the MAS near-IR reflectance ratios. Water surface emissivities at several MIR frequencies are obtained in the process of the W retrieval from several flights over the Midwest lakes. These estimated emissivities compared favorably with values calculated for a calm water surface, which are based on a di-electric permittivity model and MAS-measured surface temperatures. The results from all comparisons strongly demonstrate the soundness of the technique for estimating W.     

Electromigration-induced damage in bamboo Al interconnects

Depletion and hillock formation were examined in-situ in a scanning electron microscope (SEM) during electromigration of bamboo Al interconnect segments. Hillocks formed directly at the anode ends of the segments by epitaxial addition of Al at the bottom Al/TiN interface. Depletion occurred nonuniformly from the cathode end and stopped once the distance between the leading void and the hillock reached the critical length for electromigration at the given current density. A modified equation for the drift velocity is proposed, which includes the effect of nonuniform depletion and predicts that interconnects with nonuniform depletion are more reliable than those with uniform depletion.     

Use of the Brief Psychiatric Rating Scale to measure success in a psychosocial day program

This study contrasted six subscales of the Brief Psychiatric Rating Scale (BPRS) to determine their sensitivity to psychosocial treatment outcome. An expanded version of the BPRS was administered to 216 clients on admission to a day program. The subscale measuring hostility and suspiciousness discriminated at intake clients who were therapeutically discharged from clients who did not complete the program and predicted discharge status after the investigators controlled for the effects of demographic variables. Significant reductions in scores were obtained on five subscales for a subset of clients to whom the BPRS was readministered before discharge. The results support the use of the expanded BPRS as an evaluative tool in psychosocial rehabilitation programs.     

Effect of hygiene measures, water sanitation and oral rehydration therapy on diarrhea in children less than five years old in the south of Ivory Coast

BACKGROUND: Gastric sucrose permeability is a noninvasive marker that reliably increases in association with gastrointestinal injury due to use of nonsteroidal antiinflammatory drugs. Despite the effect of Helicobacter pylori infection on the gastric mucosa, in a previous study we were unable to demonstrate that H. pylori infection was associated with abnormal gastric sucrose permeability. Our goal in this study was to explore further whether H. pylori infection changed gastric permeability; therefore, we evaluated the effect of treatment of H. pylori infection on gastric permeability to sucrose and the relation of sucrose permeability to density of polymorphonuclear leukocytes. MATERIALS AND METHODS: Five hundred milliliters of a solution containing 100 gm of sucrose was ingested by the subject at bedtime. Overnight urine was collected and assayed for sucrose by high-performance liquid chromatography. Sucrose permeability was assessed both before and approximately 4 weeks after anti-H. pylori therapy. RESULTS: Seventeen asymptomatic H. pylori-infected volunteers participated; 8 were cured. Sucrose permeability was in the range commonly found in normal controls both before and after anti-H. pylori therapy (mean excretion, 76.3 mg; range, 13-171 mg). Gastric sucrose permeability correlated with the density of polymorphonulcear cell infiltration of the mucosa. Cure of the H. pylori infection was associated with a small but significant decrease in sucrose permeability (98.8 +/- 18 mg to 51.7 +/- 9.8 mg (p = .01). Sucrose permeability was greater in those with a high density of mucosal polymorphonuclear cells compared to those with lower scores (119.5 +/- 4 vs 71.4 +/- 13 for those with scores > or = 5 compared to scores < or = 4; p = .023). Failed therapy resulted in an increase in the mucosal density of polymorphonuclear infiltration and sucrose permeability (56.4 +/- 13 mg-99.7 +/- 19 mg pretreatment vs posttreatment, respectively; p = .031). CONCLUSION: H. pylori gastritis causes a small but measurable increase in gastric permeability to sucrose that may reflect epithelial transmigration of neutrophils.