Variational studies and replica symmetry breaking in the generalization problem of the binary perceptron

We analyze the average performance of a general class of learning algorithms for the nondeterministic polynomial time complete problem of rule extraction by a binary perceptron. The examples are generated by a rule implemented by a teacher network of similar architecture. A variational approach is used in trying to identify the potential energy that leads to the largest generalization in the thermodynamic limit. We restrict our search to algorithms that always satisfy the binary constraints. A replica symmetric ansatz leads to a learning algorithm which presents a phase transition in violation of an information theoretical bound. Stability analysis shows that this is due to a failure of the replica symmetric ansatz and the first step of replica symmetry breaking (RSB) is studied. The variational method does not determine a unique potential but it allows construction of a class with a unique minimum within each first order valley. Members of this class improve on the performance of Gibbs algorithm but fail to reach the Bayesian limit in the low generalization phase. They even fail to reach the performance of the best binary, an optimal clipping of the barycenter of version space. We find a trade-off between a good low performance and early onset of perfect generalization. Although the RSB may be locally stable we discuss the possibility that it fails to be the correct saddle point globally. ©2000 The American Physical Society.

Quantum vacuum in hot nuclear matter: A non-perturbative treatment

We derive the equation of state for hot nuclear matter using the Walecka model in a non-perturbative formalism. We include here the vacuum polarization effects arising from the nucleon and scalar mesons through a realignment of the vacuum. A ground state structure with baryon-antibaryon condensates yields the results obtained through the relativistic Hartree approximation of summing baryonic tadpole diagrams. Generalization of such a state to include the quantum effects for the scalar meson fields through the σ -meson condensates amounts to summing over a class of multiloop diagrams. The techniques of the thermofield dynamics method are used for the finite-temperature and finite-density calculations. The in-medium nucleon and sigma meson masses are also calculated in a self-consistent manner. We examine the liquid-gas phase transition at low temperatures (≈ 20 MeV), as well as apply the formalism to high temperatures to examine a possible chiral symmetry restoration phase transition.

Preparation and characterization of spherical yttrium iron garnet via coprecipitation

The purpose of this work is to obtain spherical particles yttrium iron garnet (YIG) by coprecipitation technique. The spherical particles were obtained from either nitrate or chloride salt solutions by controlling the precipitation medium. Different agents of dispersion such as PVP and ammonium iron sulfate were used to optimize the shape and size of YIG. Samples were characterized by X-ray diffraction, scanning electron microscopy and vibrating sample magnetometry. The results show that the samples phase transition takes place at 850°C (orthorhombic phase) and at 1200°C (cubic phase). Spherical shape particles, with diameter of around 0.5 μm, present magnetization values close to the bulk value (26 emu g -1). © 2001 Elsevier Science B.V. All rights reserved.

Caracterização textural e estrutural de V 2O 5/TiO 2 obtidos via sol-gel: Comparação entre secagem convencional e supercrítica

This work describes a modified sol-gel method for the preparation of V 2O 5/TiO 2 catalysts. The samples have been characterized by N 2 adsorption at 77K, x-ray diffractometry (XRD) and Fourier Transform Infrared (FT-IR). The surface area increases with the vanadia loading from 24 m 2 g -1, for pure TiO 2, to 87 m 2 g -1 for 9wt.% of V 2O 5. The rutile form is predominant for pure TiO 2 but became enriched with anatase phase when vanadia loading is increased. No crystalline V 2O 5 phase was observed in the catalysts diffractograms. Two species of surface vanadium observed by FT-IR spectroscopy a monomeric vanadyl and polymeric vanadates, the vanadyl/vanadate ratio remains practically constant.

fac-[RuCl3(no)(dppb)] (I) and mer-[RuCl3(NO)(diop)] (II) complexes: Syntheses, characterization and x-ray structures

The fac-[RuCl3(NO)(dppb)] complex I has been prepared from solution of the correspondent mer isomer in refluxing methanol (dppb = 1,4-bis(diphenylphosphino)butane). The mer-[RuCl3(NO)(diop)] (II) has been obtained from the mer-[RuCl3(diop)(H2O)] by bubbling NO for 1 h in dichloromethane (diop = 2S,3S-O-isopropylidene-2,3-dihydroxy-1,4-bis(diphenylphosphino)butane). The complexes have been characterized by microanalysis, cyclic voltammetry (CV), IR and 31P{1H} NMR spectroscopies. The crystal and molecular structures of these two compounds have been determined from X-ray studies. The mer-[RuCl3(NO)(dppb)] isomer III was characterized in solution by NMR spectra (31P{1H}, 1H{31P}, 31P-1H HETCORR, COSY 1H-1H, HMQC 1H-13C and HMBC 1H-13C). © 2002 Elsevier Science Ltd. All rights reserved.

Interfacial concentrations of chloride and bromide and selectivity for ion exchange in vesicles prepared with dioctadecyldimethylammonium halides, lipids, and their mixtures

The local concentrations of chloride, Cl b, and bromide, Br b, in the interface of vesicles prepared with dioctadecyldimethylammonium chloride, DODAC, or bromide, DODAB, dipalmitoylphosphatidylcholine, DPPC, dimyristoylphosphatidylcholine, DMPC, and mixtures of DMPC, DPPC, and DODAC were determined by chemical trapping by analyzing product yields from spontaneous dediazoniation of vesicle-bound 2,6-dimethyl-4-hexadecylbenzenediazonium ion. The values of Cl b and Br b in DODAC and DODAB vesicles increase with vesicle size, in agreement with previous data showing that counterion dissociation decreases with vesicle size. Addition of tetramethylammonium chloride displaces bromide from the DODAB vesicular interface. The value for the selectivity constant for Br/Cl exchange at the DODAB vesicular interface obtained by chemical trapping was ∼2.0, well within values obtained for comparable amphiphiles. In vesicles of DPPC the values of Cl b were very sensitive to the nature of the cation and decreased in the order Ca 2+ > Mg 2+ > Li + > Na + > K + = Cs + = Rb + ≥ +. The effect of the cation becomes more important as temperature increases above the phase transition temperature, T m, of the lipid. The values of Cl b increased sigmoidally with the mol % of DODAC in vesicles prepared with DODAC/lipid mixtures. In sonicated vesicles prepared with DODAC and DMPC (or DPPC), the values of Cl b reach local concentrations measured for the pure amphiphile at 80 mol % DODAC. These results represent the first extensive study of local concentration of ions determined directly by chemical trapping in vesicles prepared with lipids, synthetic ampliiphiles, and their mixtures.

Warm asymmetric matter in the quark-meson coupling model

In this work we study the warm equation of state of asymmetric nuclear matter in the quark-meson coupling model which incorporates explicitly quark degrees of freedom, with quarks coupled to scalar, vector, and isovector mesons. Mechanical and chemical instabilities are discussed as a function of density and isospin asymmetry. The binodal section, essential in the study of the liquid-gas phase transition is also constructed and discussed. The main results for the equation of state are compared with two common parametrizations used in the nonlinear Walecka model and the differences are outlined.

Cytotoxicity of denture base resins: Effect of water bath and microwave postpolymerization heat treatments

Purpose: This study compared the effect of two postpolymerization heat treatments on the cytotoxicity of three denture base resins on L929 cells using 3H-thymidine incorporation and MTT assays. Materials and Methods: Sample disks of Lucitone 550, QC 20, and Acron MC resins were fabricated under aseptic conditions and stored in distilled water at 37°C for 48 hours. Specimens were then divided into three groups: (1) heat treated in microwave oven for 3 minutes at 500 W; (2) heat treated in water bath at 55°C for 60 minutes; and (3) no heat treatment. Eluates were prepared by placing three disks into a sterile glass vial with 9 mL of Eagle's medium and incubating at 37°C for 24 hours. The cytotoxic effect from the eluates was evaluated using the 3H-thymidine incorporation and MTT assays, which reflect DNA synthesis levels and cell metabolism, respectively. Results: The components leached from the resins were cytotoxic to L929 cells when 3H- thymidine incorporation assay was employed. In contrast, eluates from all resins revealed noncytotoxic effects as measured by MTT assay. For both MTT assay and 3H-thymidine incorporation, the heat treatments did not decrease the cytotoxicity of the materials tested. Conclusion: Resins were graded by 3H-thymidine incorporation assay as slightly cytotoxic and by MTT assay as noncytotoxic. Cytotoxicity of the denture base materials was not influenced by microwave or water bath heat treatment.

Conformational basis for the biological activity of TOAC-labeled angiotensin II and bradykinin: Electron paramagnetic resonance, circular dichroism, and fluorescence studies

N-Terminally and internally labeled analogues of the hormones angiotensin (AII, DRVYIHPF) and bradykinin (BK, RPPGFSPFR) were synthesized containing the paramagnetic amino acid 2,2,6,6-tetramethylpiperidine-1-oxyl-4-amino-4- carboxylic acid (TOAC). TOAC replaced Asp 1 (TOAC 1-AII) and Val 3 (TOAC 3-AII) in AII and was inserted prior to Arg 1 (TOAC 0-BK) and replacing Pro 3 (TOAC 3-BK) in BK. The peptide conformational properties were examined as a function of trifluoroethanol (TFE) content and pH. Electron paramagnetic resonance spectra were sensitive to both variables and showed that internally labeled analogues yielded rotational correlation times (TC) considerably larger than N-terminally labeled ones, evincing the greater freedom of motion of the N-terminus. In TFE, τ C increased due to viscosity effects. Calculation of τ Cpeptide/τ CTOAC ratios indicated that the peptides acquired more folded conformations. Circular dichroism spectra showed that, except for TOAC 1-AII in TFE, the N-terminally labeled analogues displayed a conformational behavior similar to that of the parent peptides. In contrast, under all conditions, the TOAC 3 derivatives acquired more restricted conformations. Fluorescence spectra of All and its derivatives were especially sensitive to the ionization of Tyr 4. Fluorescence quenching by the nitroxide moiety was much more pronounced for TOAC 3-AII The conformational behavior of the TOAC derivatives bears excellent correlation with their biological activity, since, while the N-terminally labeled peptides were partially active, their internally labeled counterparts were inactive [Nakaie, C. R., et al., Peptides 2002, 23, 65-70]. The data demonstrate that insertion of TOAC in the middle of the peptide chain induces conformational restrictions that lead to loss of backbone flexibility, not allowing the peptides to acquire their receptor-bound conformation. © 2004 Wiley Periodicals, Inc.

Quantum topology change and large-N gauge theories

We study a model for dynamical localization of topology using ideas from non-commutative geometry and topology in quantum mechanics. We consider a collection X of N one-dimensional manifolds and the corresponding set of boundary conditions (self-adjoint extensions) of the Dirac operator D. The set of boundary conditions encodes the topology and is parameterized by unitary matrices g. A particular geometry is described by a spectral triple x(g) = (A X, script H sign X, D(g)). We define a partition function for the sum over all g. In this model topology fluctuates but the dimension is kept fixed. We use the spectral principle to obtain an action for the set of boundary conditions. Together with invariance principles the procedure fixes the partition function for fluctuating topologies. The model has one free-parameter β and it is equivalent to a one plaquette gauge theory. We argue that topology becomes localized at β = ∞ for any value of N. Moreover, the system undergoes a third-order phase transition at β = 1 for large-N. We give a topological interpretation of the phase transition by looking how it affects the topology. © SISSA/ISAS 2004.

Spinodal decomposition in pure-gauge QCD

Spinodal decomposition in a model of pure-gauge SU(2) theory that incorporates a deconfinement phase transition is investigated by means of real-time lattice simulations of the fully nonlinear Ginzburg-Landau equation. Results are compared with a Glauber dynamical evolution using Monte Carlo simulations of pure-gauge lattice QCD. © 2005 American Institute of Physics.

SEM analysis of internal adaptation of adhesive restorations after contamination with saliva

Purpose: This study tested the null hypothesis that different treatments of saliva-contaminated substrate would not affect microgap formation at the dentin walls of bonded restorations. Materials and Methods: Forty freshly extracted human molars received standardized Class V preparations on buccal and lingual surfaces. The specimens were assigned to four experimental groups (n = 20): [G1] no contamination (control group), [G2] saliva contamination (10 s) after etching followed by 5 s air stream; [G3] saliva contamination after etching and rinsed for 10 s; and [G4] re-etching for 10 s after saliva contamination. All specimens were restored with a one-bottle adhesive (Single Bond, 3M ESPE) and microhybrid composite resin (Filtek Z250, 3M ESPE) according to the manufacturer's instructions. The specimens were thermocycled, sectioned through the center of the restoration, and then processed for SEM. Microgaps were measured at the axial wall at 1500X magnification. The data were submitted to Kruskal-Wallis nonparametric statistical analysis at p < 0.05. Results: The data revealed that different groups resulted in a statistically significant difference (p < 0.01) in gap formation. Air drying [G2] and rinsing [G3] the saliva-contaminated dentin resulted in similar microgap values (p > 0.05). However, re-etching the dentin after saliva contamination [G4] increased microgap formation (p < 0.05) when compared with the groups G1 and G2. Although air drying and rinsing produced results comparable to noncontaminated dentin, the presence of microgaps was not completely eliminated. Conclusion: Contaminated saliva did not prevent hybrid layer formation; however, it did reduce the adaptation of the restorative material to bonded surfaces.

Adsorption of silanes bearing nitrogenated Lewis bases on SiO 2/Si (100) model surfaces

The present paper describes the one-pot procedure for the formation of self-assembled thin films of two silanes on the model oxidized silicon wafer, SiO2/Si. SiO2/Si is a model system for other surfaces, such as glass, quartz, aerosol, and silica gel. MALDI-TOF MS with and without a matrix, XPS, and AFM have confirmed the formation of self-assembled thin films of both 3-imidazolylpropyltrimethoxysilane (3-IPTS) and 4-(N- propyltriethoxysilane-imino)pyridine (4-PTSIP) on the SiO2/Si surface after 30 min. Longer adsorption times lead to the deposition of nonreacted 3-IPTS precursors and the formation of agglomerates on the 3-IPTS monolayer. The formation of 4-PTSIP self-assembled layers on SiO2/Si is also demonstrated. The present results for the flat SiO2/Si surface can lead to a better understanding of the formation of a stationary phase for affinity chromatography as well as transition-metal-supported catalysts on silica and their relationship with surface roughness and ordering. The 3-IPTS and 4-PTSIP modified SiO2/Si wafers can also be envisaged as possible built-on-silicon thin-layer chromatography (TLC) extraction devices for metal determination or N-heterocycle analytes, such as histidine and histamine, with on-spot MALDI-TOF MS detection. © 2005 Elsevier Inc. All rights reserved.

Anelastic spectroscopy in superconducting oxides

Since the discovery of the high Tc superconductors, several works have been made about the different properties of these materials. Anelastic spectroscopy experiments are sensitive tools to the study of defects in solids and phase transitions. By this technique, we can distinguish the different types of atomic jumps that happen to different temperatures. The intensity of the peaks in the anelastic spectrum and the step in the torsional modulus are related with the concentration of the relaxing entities, and the position of the peaks is determined by its mobility. In this paper, the study on Bi and Sm based superconducting oxides was made by anelastic relaxation measurements using a torsion pendulum. The samples were submitted to successive thermal treatments in high vacuum, in the temperature range between 100 K and 650 K, heating rate about 1 K/min. For Bi based superconducting oxides the results shown two peaks, that were associated to interstitial oxygen mobility and to orthorhombic to monoclinic phase transition. For Sm based superconducting oxides the results shown a relaxation peak that was attributed to the jumps of the oxygen atoms in the inter-chains O1 and 05 of the lattice.

Cytotoxicity of hard chairside reline resins: Effect of microwave irradiation and water bath postpolymerization treatments

Purpose: To evaluate the influence of water bath and microwave postpolymerization treatments on the cytotoxicity of 6 hard reline acrylic resins. Materials and Methods: The materials tested were Tokuso Rebase Fast (TR), Ufi Gel Hard (UGH), Duraliner II (D), Kooliner (K), New Truliner (NT), and Light Liner (LL). LL resin was additionally tested with an air-barrier coating (LLABC). Nine disks of each material (10 × 1 mm) were made and divided into 3 groups: group 1 (no postpolymerization treatment); group 2 (postpolymerization in microwave oven); group 3 (postpolymerization in water bath at 55°C for 10 minutes). L929 cells were cultured in 96-well plates and incubated for 24 hours in Eagle's medium. Eluates prepared from the disks or medium without disks (control) replaced the medium. Cytotoxicity was assessed by both dehydrogenase succinic activity (MTT) assay and incorporation of radioactive 3H-thymidine assay. Tests were carried out in quadruplicate and repeated twice. Differences between groups were determined by analysis of variance with Tukey multiple-comparison intervals (α = .05). Results: For MTT assay, the postpolymerization treatments had no effect on the cytotoxicity of all materials (P > .05). For 3H-thymidine assay, the postpolymerization treatments significantly decreased the cytotoxicity of UGH (P < .05). The cytotoxicity of K, NT, LL, and LLABC increased after microwave irradiation (P < .05). TR, NT, and LLABC showed an increase in cytotoxicity after water bath (P < .05). Conclusion: When assessed by MTT assay, the cytotoxicity of the materials was not affected by postpolymerization treatments. 3H-Thymidine assay showed that the cytotoxicity of the resins was not improved by the postpolymerization treatments, with the exception of UGH.