Classification calibration dimension for general multiclass losses

We study consistency properties of surrogate loss functions for general multiclass classification problems, defined by a general loss matrix. We extend the notion of classification calibration, which has been studied for binary and multiclass 0-1 classification problems (and for certain other specific learning problems), to the general multiclass setting, and derive necessary and sufficient conditions for a surrogate loss to be classification calibrated with respect to a loss matrix in this setting. We then introduce the notion of \emph{classification calibration dimension} of a multiclass loss matrix, which measures the smallest `size' of a prediction space for which it is possible to design a convex surrogate that is classification calibrated with respect to the loss matrix. We derive both upper and lower bounds on this quantity, and use these results to analyze various loss matrices. In particular, as one application, we provide a different route from the recent result of Duchi et al.\ (2010) for analyzing the difficulty of designing `low-dimensional' convex surrogates that are consistent with respect to pairwise subset ranking losses. We anticipate the classification calibration dimension may prove to be a useful tool in the study and design of surrogate losses for general multiclass learning problems.

Total Synthesis of (+)-Seimatopolide A

Total synthesis of 10-membered lactone (+)-seimatopolide A is presented from furfural. Key reactions in the present strategy include the effective use of furan as a E-but-2-ene-1,4-dione surrogate, Nagao acetate aldol reaction, and Shiina lactonization.

Give-and-take based peer-to-peer content distribution networks

Content Distribution Networks (CDNs) are widely used to distribute data to large number of users. Traditionally, content is being replicated among a number of surrogate servers, leading to high operational costs. In this context, Peer-to-Peer (P2P) CDNs have emerged as a viable alternative. An issue of concern in P2P networks is that of free riders, i.e., selfish peers who download files and leave without uploading anything in return. Free riding must be discouraged. In this paper, we propose a criterion, the Give-and-Take (G&T) criterion, that disallows free riders. Incorporating the G&T criterion in our model, we study a problem that arises naturally when a new peer enters the system: viz., the problem of downloading a `universe' of segments, scattered among other peers, at low cost. We analyse this hard problem, and characterize the optimal download cost under the G&T criterion. We propose an optimal algorithm, and provide a sub-optimal algorithm that is nearly optimal, but runs much more quickly; this provides an attractive balance between running time and performance. Finally, we compare the performance of our algorithms with that of a few existing P2P downloading strategies in use. We also study the computation time for prescribing the strategy for initial segment and peer selection for the newly arrived peer for various existing and proposed algorithms, and quantify cost-computation time trade-offs.

Effect of viscosity and surface tension on breakup and coalescence of bicomponent sprays

This work deals with an experimental study of the breakup characteristics of liquids with different surface tension and viscosity from a hollow cone hydraulic injector nozzle induced by pressure-swirl. The experiments were conducted at Reynolds numbers Re-p=9500-23,000. The surface tension and viscosity of the surrogate fuels were altered from 72 to 30 mN/m and 1.1 to 1.6 mN s/m(2), respectively. High speed photography and Phase Doppler Particle Anemometry were utilized to study the atomization process. Velocity and drop size measurements of the spray using PDPA in both axial and radial directions indicate a dependency on surface tension. However, these effects are dominant only at low Reynolds numbers and are negligible at high Reynolds number. Downstream of the nozzle, coalescence of droplets due to collision was also found to be significant and the diameters were compared for different liquids. For viscous fluids up to 1.6 cP, the independent effects of viscosity and injection pressure are studied. In general, the spray cone angle increases with increase in pressure. At high pressures, an increase in viscosity leads to higher drop sizes following primary and secondary breakup compared to water. This study will extend our understanding of surrogate fuel film breakup and highlight the importance of long and short wavelength instabilities. (C) 2013 Elsevier Ltd. All rights reserved

A Mycobacterial Phosphoribosyltransferase Promotes Bacillary Survival by Inhibiting Oxidative Stress and Autophagy Pathways in Macrophages and Zebrafish

Mycobacterium tuberculosis employs various strategies to modulate host immune responses to facilitate its persistence in macrophages. The M. tuberculosis cell wall contains numerous glycoproteins with unknown roles in pathogenesis. Here, by using Concanavalin A and LC-MS analysis, we identified a novel mannosylated glycoprotein phosphoribosyltransferase, encoded by Rv3242c from M. tuberculosis cell walls. Homology modeling, bioinformatic analyses, and an assay of phosphoribosyltransferase activity in Mycobacterium smegmatis expressing recombinant Rv3242c (MsmRv3242c) confirmed the mass spectrometry data. Using Mycobacterium marinum-zebrafish and the surrogate MsmRv3242c infection models, we proved that phosphoribosyltransferase is involved in mycobacterial virulence. Histological and infection assays showed that the M. marinum mimG mutant, an Rv3242c orthologue in a pathogenic M. marinum strain, was strongly attenuated in adult zebrafish and also survived less in macrophages. In contrast, infection with wild type and the complemented Delta mimG: Rv3242c M. marinum strains showed prominent pathological features, such as severe emaciation, skin lesions, hemorrhaging, and more zebrafish death. Similarly, recombinant Msm Rv3242c bacteria showed increased invasion in non-phagocytic epithelial cells and longer intracellular survival in macrophages as compared with wild type and vector control M. smegmatis strains. Further mechanistic studies revealed that the Rv3242c- and mimG-mediated enhancement of intramacrophagic survival was due to inhibition of autophagy, reactive oxygen species, and reduced activities of superoxide dismutase and catalase enzymes. Infection with MsmRv3242c also activated the MAPK pathway, NF-kappa B, and inflammatory cytokines. In summary, we show that a novel mycobacterial mannosylated phosphoribosyltransferase acts as a virulence and immunomodulatory factor, suggesting that it may constitute a novel target for antimycobacterial drugs.

THE EFFECTS OF SURFACTANT ON SIMPLEX NOZZLE SPRAY BEHAVIOR AND ITS COMPARISON TO LIQUID FUELS

Pressure-swirl nozzles (simplex nozzles) are used in various field applications such as aero-engines, power generation, spray painting and agricultural irrigation. For this particular nozzle, research in the past decade has dealt with the development of numerical models for predicting droplet distribution profiles. Although these results have been valuable, the experimental results have been contradictory, therefore fundamental understanding of the influence of properties in nozzle is important. This paper experimentally investigates the effect of surfactants on breakup and coalescence. Since most of the fuels and biofuels have low surface tension compared to water, a comparative analysis between a surfactant solution and a liquid fuel is imperative. For this experimental study, a simplex nozzle characterized as flow number 0.4 will be utilized. The injection pressures will range from 0.3 - 4Mpa while altering the surface tension from 72 to 28mN/m. By applying Phase Doppler Particle Anemometry (PDPA) which is a non-intrusive laser diagnostic technique, the differences in spray characteristics due to spray surface tension can be highlighted. The average droplet diameter decreases for a low surface tension fluid in the axial direction in comparison to pure water. The average velocity of droplets is surprisingly lower in the same spray zone. Measurements made in the radial direction show no significant changes, but at the locations close to the nozzle, water droplets have larger diameter and velocity. The results indicate the breakup and coalescence regimes have been altered when surface tension is lowered. A decrease in surface tension alters the breakup length while increasing the spray angle. Moreover, higher injection pressure shortens the breakup length and decrease in overall diameter of the droplets. By performing this experimental study the fundamentals of spray dynamics, such as spray formation, liquid breakup length, and droplet breakup regimes can be observed as a function of surface tension and how a surrogate fuel compares with a real fuel for experimental purposes. This knowledge potentially will lead to designing a better atomizer or new biofuels.

Experimental studies on evaporation of fuel droplets under forced convection using spray in crossflow methodology

Experimental data on evaporation of droplets of decane, Jet-A1, and Jet-A1 surrogate are generated using a spray in crossflow configuration. The advantage of a crossflow configuration is that it enables us to study droplet evaporation under forced convective conditions involving droplet diameters of size relevant in practical combustors. Specifically, spray from an airblast atomizer is injected into a preheated crossflow of air and the resulting spray is characterized in terms of spray structure along with droplet size and velocity. An existing correlation for the spray trajectory is modified to incorporate the effect of elevated temperature, and is found to be in good agreement with the experimental data. Droplet sizes and velocities are measured at different locations along the crossflow direction to assess droplet evaporation. Specifically, droplets having size less than 25-mu m are selected for further analysis since these droplets are observed to exhibit velocities which are aligned with the crossflow. By comparing the droplet diameter profiles at upstream and downstream locations, the evaporation constant k for the d(2)-law is obtained iteratively. To assess the efficacy of the values of k obtained, the calculated droplet size distribution using the proposed k values at the downstream location is compared with the measured droplet size distribution at that location. A reasonably good match is found for all the three liquids confirming the validity of the analysis. (C) 2015 Elsevier Ltd. All rights reserved.

Systematic Analysis of Mycobacterial Acylation Reveals First Example of Acylation-mediated Regulation of Enzyme Activity of a Bacterial Phosphatase

Protein lysine acetylation is known to regulate multiple aspects of bacterial metabolism. However, its presence in mycobacterial signal transduction and virulence-associated proteins has not been studied. In this study, analysis of mycobacterial proteins from different cellular fractions indicated dynamic and widespread occurrence of lysine acetylation. Mycobacterium tuberculosis proteins regulating diverse physiological processes were then selected and expressed in the surrogate host Mycobacterium smegmatis. The purified proteins were analyzed for the presence of lysine acetylation, leading to the identification of 24 acetylated proteins. In addition, novel lysine succinylation and propionylation events were found to co-occur with acetylation on several proteins. Protein-tyrosine phosphatase B (PtpB), a secretory phosphatase that regulates phosphorylation of host proteins and plays a critical role in Mycobacterium infection, is modified by acetylation and succinylation at Lys-224. This residue is situated in a lid region that covers the enzyme's active site. Consequently, acetylation and succinylation negatively regulate the activity of PtpB.

A Method for Delineation of Bone Surfaces in Photoacoustic Computed Tomography of the Finger

Photoacoustic (PA) imaging of interphalangeal peripheral joints is of interest in the context of using the synovial membrane as a surrogate marker of rheumatoid arthritis. Previous work has shown that ultrasound (US) produced by absorption of light at the epidermis reflects on the bone surfaces within the finger. When the reflected signals are backprojected in the region of interest, artifacts are produced, confounding interpretation of the images. In this work, we present an approach where the PA signals known to originate from the epidermis are treated as virtual US transmitters, and a separate reconstruction is performed as in US reflection imaging. This allows us to identify the bone surfaces. Furthermore, the identification of the joint space is important as this provides a landmark to localize a region-of-interest in seeking the inflamed synovial membrane. The ability to delineate bone surfaces allows us to identify not only the artifacts but also the interphalangeal joint space without recourse to new US hardware or a new measurement. We test the approach on phantoms and on a healthy human finger.

IMAGE DENOISING USING OPTIMALLY WEIGHTED BILATERAL FILTERS: A SURE AND FAST APPROACH

The bilateral filter is known to be quite effective in denoising images corrupted with small dosages of additive Gaussian noise. The denoising performance of the filter, however, is known to degrade quickly with the increase in noise level. Several adaptations of the filter have been proposed in the literature to address this shortcoming, but often at a substantial computational overhead. In this paper, we report a simple pre-processing step that can substantially improve the denoising performance of the bilateral filter, at almost no additional cost. The modified filter is designed to be robust at large noise levels, and often tends to perform poorly below a certain noise threshold. To get the best of the original and the modified filter, we propose to combine them in a weighted fashion, where the weights are chosen to minimize (a surrogate of) the oracle mean-squared-error (MSE). The optimally-weighted filter is thus guaranteed to perform better than either of the component filters in terms of the MSE, at all noise levels. We also provide a fast algorithm for the weighted filtering. Visual and quantitative denoising results on standard test images are reported which demonstrate that the improvement over the original filter is significant both visually and in terms of PSNR. Moreover, the denoising performance of the optimally-weighted bilateral filter is competitive with the computation-intensive non-local means filter.