Evolutionary optimization for ranking how-to questions based on user-generated contents

In this work, a new evolutionary model is proposed for ranking answers to non-factoid (how-to) questions in community question-answering platforms. The approach combines evolutionary computation techniques and clustering methods to effectively rate best answers from web-based user-generated contents, so as to generate new rankings of answers. Discovered clusters contain semantically related triplets representing question–answers pairs in terms of subject-verb-object, which is hypothesized to improve the ranking of candidate answers. Experiments were conducted using our evolutionary model and concept clustering operating on large-scale data extracted from Yahoo! Answers. Results show the promise of the approach to effectively discovering semantically similar questions and improving the ranking as compared to state-of-the-art methods.     

Use of CPWL approximations in the design of a numerical nonlinearregulator

This paper presents the state regulation problem for nonlinear plants with initial conditions in a given bounded region, proposed originally by Kreisselmeier and Birkholzer (1994). An efficient computer implementable algorithm is presented, based on canonical piecewise linear approximation for the discrete model of the plant     

Electrochemical behaviour of chalcopyrite in the presence of silver and Sulfolobus bacteria

The electrochemical behaviour of massive chalcopyrite electrodes has been studied in an acid medium (pH1.5) containing silver ions (0.02gdm?3Ag+) and thermophilic bacteria (68°C). Preliminary tests on particulate electrodes made from graphite, elemental sulfur and Ag2S were included to determine the electrochemical response of reactants (Ag+) and products (S° and Ag2S) associated with the dissolution of chalcopyrite in the presence of silver. Massive chalcopyrite electrodes under potential scan showed a dependence on the dissolution of the Ag2S film with both the time of contact with the silver solution and [Ag+]. As well as Ag2S, metallic silver was detected on the chalcopyrite surface. It has been demonstrated that Fe3+ and bacteria play an important role in the regeneration of the Ag2S film. The breakdown of this film is a requirement for the further dissolution of chalcopyrite. The bioleaching of chalcopyrite with thermophilic microorganisms in the presence of silver decreased the decomposition potential of the electrode and favoured its electrodissolution. Bioleaching treatment in the presence of silver ions for periods of time longer than two weeks did not improve the surface reactivity. However, in the initial stages of the process, the lower reactivity of the bioleached electrodes was probably related to a toxic effect of silver on the microorganisms.     

Morphological features and melting behavior of nanocomposites based on isotactic polypropylene and multiwalled carbon nanotubes

Nanocomposites based on low molar mass isotactic polypropylene (iPP) and a low concentrations (1–2 wt %) of multiwalled carbon nanotubes (MWCNTs) were studied using thermal analysis, optical and electronic microscopy, and X‐ray diffraction/scattering techniques. It was first determined that MWCNT decrease induction time and act as nucleating agents of the iPP crystals during nonisothermal crystallization. One of the consequences of the nucleation effect was that the original spherulitic morphology of iPP was transformed into a fibrillar‐like. The corresponding long period of the original well‐defined lamellar structure slightly increased suggesting the formation of thicker crystals in samples containing MWCNT. The nature of the α‐iPP crystalline structure was not affected by MWCNT. After nonisothermal crystallization, two melting endotherms were present during thermal scanning of the iPP/MWCNT nanocomposites their proportion changing with the heating rate. After resolving the total DSC signal in its components using MDSC, the overall evolution of such behavior could be explained in terms of the melting/recrystallization mechanism. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007     

Three-dimensional digitization of highly reflective and transparent objects using multi-wavelength range sensing

Digitizing specular and transparent objects pose significant problems using traditional 3-D scanning techniques due to the reflection and refraction that interfere with the optical scanning process used for triangulation. In this paper, we present how one can digitize those difficult objects by modifying a commercial 3-D acquisition system with an interchangeable ultraviolet and infrared light source. Experimental results show that the proposed technique can generate accurate 3-D models of these optically challenging objects without major modifications to the 3-D scanner. The results were obtained without preprocessing or multi-view manipulations. The precision of the 3-D measurements is evaluated relative to the visible spectrum acquisition obtained by painting the test objects with matte paint to suppress optical difficulties. Results shows that wavelength changes in the 3-D acquisition system do not change the scanner precision but solve many of the issues that specular and transparent objects poses.     

Influence of molybdenum ion implantation on the localized corrosion resistance of a high strength aluminium alloy

The effect of Mo⁺ implantation on corrosion properties of AA 7075 (T6 and T73) has been studied. The induced modifications in the surfaces have been analysed by XPS and GAXRD. EXCO tests and electrochemical DC and AC techniques have been used to characterize the localized corrosion resistance. The effect of implantation is highly dependent on the ageing conditions. A marked increasing of the impedance of T6 samples after implantation has been measured, while the effect of Mo on T73 samples is less clear. Impedance spectra allow analysing the differences in the evolution of localized corrosion between both tempers.     

Polycatechol Nanoparticle MRI Contrast Agents

Amphiphilic triblock copolymers containing Fe^III‐catecholate complexes formulated as spherical‐ or cylindrical‐shaped micellar nanoparticles ( SMN and CMN , respectively) are described as new T ₁‐weighted agents with high relaxivity, low cytotoxicity, and long‐term stability in biological fluids. Relaxivities of both SMN and CMN exceed those of established gadolinium chelates across a wide range of magnetic field strengths. Interestingly, shape‐dependent behavior is observed in terms of the particles' interactions with HeLa cells, with CMN exhibiting enhanced uptake and contrast via magnetic resonance imaging (MRI) compared with SMN . These results suggest that control over soft nanoparticle shape will provide an avenue for optimization of particle‐based contrast agents as biodiagnostics. The polycatechol nanoparticles are proposed as suitable for preclinical investigations into their viability as gadolinium‐free, safe, and effective imaging agents for MRI contrast enhancement.     

Aerobic granulation in a mechanical stirred SBR: treatment of low organic loads

Aerobic granular sludge was produced in a sequencing batch reactor (SBR) characterized by a height to diameter ratio of 2.5 and the use of mechanical stirring. Compact and regular aerobic granules of up to 1.75 mm of average diameter were formed in the reactor with an organic loading rate of 1.75 kg COD/(m3 d). Settling properties of the obtained aggregates were: sludge volumetric index of 30-40 mL/g VSS and settling velocity higher than 8 m/h. The effects of different carbon to nitrogen ratios (TOC/N) in the feeding on the organic matter oxidation and nitrification process were studied. The concentration of organic matter in the feeding was stepwise reduced (from 190.0 to 37.5 mg TOC/L) and ammonium increased (from 25 to 50 mg NH4+ -N/L). TOC/N ratios of 7.50, 3.00, 1.50 and 0.75 g/g in the feeding were tested. The TOC removal percentage was around 80-95% during the whole operational period and the N removal percentages obtained in the reactor were up to 40%, however, physical properties of the granules were not maintained.     

Processing conditions—fracture toughness relationships of asphalt concrete mixtures

The effect of processing conditions (dynamic compaction) on the fatigue crack propagation behaviour of AC-20 asphalt concrete mixture was studied. Beams were prepared from AC-20 asphalt binder containing 8% asphalt by weight with and without dynamic compaction. The gradation used was Ohio Department of Transportation item 403, and was kept the same. Flexural static tests were conducted to determine the effect of dynamic compaction on both the ultimate strength and flexural modulus. Flexural fatigue tests were conducted on three identical notched specimens prepared using each of the two compaction techniques. Parameters controlling the crack propagation process were evaluated; namely, the energy release rate and the change in work, W _i expended on damage formation and history dependent viscous dissipation processes within the active zone (region ahead of the crack tip). The modified crack layer (MCL) model was employed to extract the specific energy of damage, , a material parameter characteristic of the asphalt concrete mixture's resistance to crack propagation, and the dissipative coefficient, . It has been found that the dynamically and statically compacted AC-20 mixture displayed superior fracture resistance, as reflected in and . Also, the ultimate strength and modulus increased by about two-fold. Scanning electron microscopic examination revealed an obvious change in the morphology of the fracture surface. This is manifested in the appearance of a finer more dense texture in the case of the dynamically and statically compacted mixture. In addition, smaller more frequent dimples in binder rich areas are indicative of better adhesion between the binder and the aggregate. This in turn contributes to the increased fracture resistance of the dynamically and statically compacted AC-20 asphalt concrete mixture.     

Effective proximity retrieval by ordering permutations

We introduce a new probabilistic proximity search algorithm for range and K-nearest neighbor (K-NN) searching in both coordinate and metric spaces. Although there exist solutions for these problems, they boil down to a linear scan when the space is intrinsically high-dimensional, as is the case in many pattern recognition tasks. This, for example, renders the K-NN approach to classification rather slow in large databases. Our novel idea is to predict closeness between elements according to how they order their distances towards a distinguished set of anchor objects. Each element in the space sorts the anchor objects from closest to farthest to it, and the similarity between orders turns out to be an excellent predictor of the closeness between the corresponding elements. We present extensive experiments comparing our method against state-of-the-art exact and approximate techniques, both in synthetic and real, metric and non-metric databases, measuring both CPU time and distance computations. The experiments demonstrate that our technique almost always improves upon the performance of alternative techniques, in some cases by a wide margin.