Structured Learning for Non-Smooth Ranking losses

Learning to rank from relevance judgment is an active research area. Itemwise score regression, pairwise preference satisfaction, and listwise structured learning are the major techniques in use. Listwise structured learning has been applied recently to optimize important non-decomposable ranking criteria like AUC (area under ROC curve) and MAP(mean average precision). We propose new, almost-lineartime algorithms to optimize for two other criteria widely used to evaluate search systems: MRR (mean reciprocal rank) and NDCG (normalized discounted cumulative gain)in the max-margin structured learning framework. We also demonstrate that, for different ranking criteria, one may need to use different feature maps. Search applications should not be optimized in favor of a single criterion, because they need to cater to a variety of queries. E.g., MRR is best for navigational queries, while NDCG is best for informational queries. A key contribution of this paper is to fold multiple ranking loss functions into a multi-criteria max-margin optimization.The result is a single, robust ranking model that is close to the best accuracy of learners trained on individual criteria. In fact, experiments over the popular LETOR and TREC data sets show that, contrary to conventional wisdom, a test criterion is often not best served by training with the same individual criterion.

Structure and bonding of MCB5H7 and its sandwiched dimer CB5H6M–MCB5H6 (M = Si, Ge, Sn): Isomer stability and preference for slip distorted structure

We find sandwiched metal dimers CB5H6M–MCB5H6 (M = Si, Ge, Sn) which are minima in the potential energy surface with a characteristic M–M single bond. The NBO analysis and the M–M distances (Å) (2.3, 2.44 and 2.81 for M = Si, Ge, Sn) indicate substantial M–M bonding. Formal generation of CB5H6M–MCB5H6 has been studied theoretically. Consecutive substitution of two boron atoms in B7H−27 by M (Si, Ge, Sn) and carbon, respectively followed by dehydrogenation may lead to our desired CB5H6M–MCB5H6. We find that the slip distorted geometry is preferred for MCB5H7 and its dehydrogenated dimer CB5H6M–MCB5H6. The slip-distortion of M–M bond in CB5H6M–MCB5H6 is more than the slip distortion of M–H bond in MCB5H7. Molecular orbital analysis has been done to understand the slip distortion. Larger M–M bending (CB5H6M–MCB5H6) in comparison with M–H bending (MCB5H7) is suspected to be encouraged by stabilization of one of the M–M π bonding MO’s. Preference of M to occupy the apex of pentagonal skeleton of MCB5H7 over its icosahedral analogue MCB10H11 has been observed.

Genetic characterization of Indian type O FMD virus 3A region in context with host cell preference

The 3A region of foot-and-mouth disease virus has been implicated in host range and virulence. For example, amino acid deletions in the porcinophilic strain (O/TAW/97) at 93-102 aa of the 153 codons long 3A protein have been recognized as the determinant of species specificity. In the present study, 18 type 0 FMDV isolates from India were adapted in different cell culture systems and the 3A sequence was analyzed. These isolates had complete 3A coding sequence (153 aa) and did not exhibit growth restriction in cells based on species of origin. The 3A region was found to be highly conserved at N-terminal half (1-75 aa) but exhibited variability or substitutions towards C-terminal region (80-153). Moreover the amino acid substitutions were more frequent in recent Indian buffalo isolates but none of the Indian isolates showed deletion in 3A protein, which may be the reason for the absence of host specificity in vitro. Further inclusive analysis of 3A region will reveal interesting facts about the variability of FMD virus 3A region in an endemic environment. (C) 2010 Elsevier B.V. All rights reserved.

A new scheme for establishing pairwise keys for wireless sensor networks

This paper addresses the problem of secure path key establishment in wireless sensor networks that uses the random key predistribution technique. Inspired by the recent proxy-based scheme in [1] and [2], we introduce a fiiend-based scheme for establishing pairwise keys securely. We show that the chances of finding friends in a neighbourhood are considerably more than that of finding proxies, leading to lower communication overhead. Further, we prove that the friendbased scheme performs better than the proxy-based scheme in terms of resilience against node capture.

Characterization and Acceptor Preference of a Soluble Meningococcal Group C Polysialyltransferase

Vaccines against Neisseria meningitidis group C are based on its alpha-2,9-linked polysialic acid capsular polysaccharide. This polysialic acid expressed on the surface of N. meningitidis and in the absence of specific antibody serves to evade host defense mechanisms. The polysialyltransferase (PST) that forms the group C polysialic acid (NmC PST) is located in the cytoplasmic membrane. Until recently, detailed characterization of bacterial polysialyltransferases has been hampered by a lack of availability of soluble enzyme preparations. We have constructed chimeras of the group C polysialyltransferase that catalyzes the formation alpha-2,9-polysialic acid as a soluble enzyme. We used site-directed mutagenesis to determine the region of the enzyme necessary for synthesis of the alpha-2,9 linkage. A chimera of NmB and NmC PSTs containing only amino acids 1 to 107 of the NmB polysialyltransferase catalyzed the synthesis of alpha-2,8-polysialic acid. The NmC polysialyltransferase requires an exogenous acceptor for catalytic activity. While it requires a minimum of a disialylated oligosaccharide to catalyze transfer, it can form high-molecular-weight alpha-2,9-polysialic acid in a nonprocessive fashion when initiated with an alpha-2,8-polysialic acid acceptor. De novo synthesis in vivo requires an endogenous acceptor. We attempted to reconstitute de novo activity of the soluble group C polysialyltransferase with membrane components. We found that an acapsular mutant with a defect in the polysialyltransferase produces outer membrane vesicles containing an acceptor for the alpha-2,9-polysialyltransferase. This acceptor is an amphipathic molecule and can be elongated to produce polysialic acid that is reactive with group C-specific antibody.

DNA Condensation by the Rat Spermatidal Protein TP2 Shows GC-Rich Sequence Preference and Is Zinc Dependent

Transition protein-2 (TP2), isolated from rat testes, was recently shown to be a zinc metalloprotein. We have now carried out a detailed analysis of the DNA condensing properties of TP2 with various polynucleotides using circular dichroism spectroscopy. The condensation of the alternating copolymers by TP2 (incubated with 10 mu M ZnSO4), namely, poly(dG-dC). poly(dG-dC) and poly(dA-dT). poly(dA-dT), was severalfold higher than condensation of either of the homoduplexes poly(dG). poly-(dC) and poly(dA). poly(dT) or rat oligonucleosomal DNA. Between the two alternating copolymers, poly(dG-dC). poly(dG-dC) was condensed 3.2-fold more effectively than poly(dA-dT). poly(dA-dT). Preincubation of TP2 with 5 mM EDTA significantly reduced its DNA-condensing property. Interestingly, condensation of the alternating copolymer poly(dI-dC). poly(dI-dC) by TP2 was much less as compared to that of poly(dG-dC). poly(dG-dC). The V8 protease-derived N-terminal fragment (88 aa) condensed poly(dA-dT). poly(dA-dT) to a very small extent but did not have any effect on poly(dG-dC). poly-(dG-dC). The C-terminal fragment (28 aa) was able to condense poly(dA-dT) . poly(dA-dT) more effectively than poly(dG-dC). poly(dG-dC). These results suggest that TP2 in its zinc-coordinated form condenses GC-rich polynucleotides much more effectively than other types of polynucleotides. Neither the N-terminal two-thirds of TP2 which is the zinc-binding domain nor the C-terminal basic domain are as effective as intact TP2 in bringing about condensation of DNA.

Study of translational and rotational mobility and orientational preference of ethane in one-dimensional channels

Structural and dynamical properties of ethane in one-dimensional channels of AlPO4-5 and carbon nanotube have been investigated at dilute concentration with the help of molecular dynamics simulation. Density distributions and orientational structure of ethane have been analyzed. Repulsive interactions seem to play an important role when ethane is located in the narrow part of the AlPO4-5 channel. In AlPO4-5, parallel orientation is predominant over perpendicular orientation except when ethane is located in the broader part of the channel. Unlike in the case of single-file diffusion, our results in carbon nanotube show that at dilute concentrations the mean squared displacement, mu(2)(t) approximate to t(alpha), alpha = 1.8. The autocorrelation function for the z-component of angular velocity of ethane in space-fixed frame of reference shows a pronounced negative correlation. This is attributed to the restriction in the movement of ethane along the x- and y- directions. It is seen that the ratio of reorientational correlation times does not follow the Debye model for confined ethane but it is closer to the predictions of the Debye model for bulk ethane.

An Efficient Scheme for Establishing Pairwise Keys for Wireless Sensor Networks

This paper addresses the problem of secure path key establishment in wireless sensor networks that uses the random key pre-distribution technique. Inspired by the recent proxy-based scheme in the work of Ling and Znati (2005) and Li et al. (2005), we introduce a friend-based scheme for establishing pairwise keys securely. We show that the chances of finding friends in a neighbourhood are considerably more than that of finding proxies, leading to lower communication overhead. Further, we prove that the friend-based scheme performs better than the proxy-based scheme both in terms of resilience against node capture as well as in energy consumption for pairwise key establishment, making our scheme more feasible.

Polymorphism and tautomeric preference in fenobam and the utility of NLO response to detect polymorphic impurities

Crystal structures of polymorphs and solvatomorphs of the potential anxiolytic drug fenobam exhibit an exclusive preference for one of the two possible tautomeric structures. A novel methodology based on nonlinear optical response has been successfully employed to detect the presence of a polymorphic impurity in a mixture of polymorphs.

How general is a female mating preference for clustered males in lekking species? a meta-analysis

A principal hypothesis for the evolution of leks (rare and intensely competitive territorial aggregations) is that leks result from females preferring to mate with clustered males. This hypothesis predicts more female visits and higher mating success per male on larger leks. Evidence for and against this hypothesis has been presented by different studies, primarily of individual populations, but its generality has not yet been formally investigated. We took a meta-analytical approach towards formally examining the generality of such a female bias in lekking species. Using available published data and using female visits as an index of female mating bias, we estimated the shape of the relationship between lek size and total female visits to a lek, female visits per lekking male and, where available, per capita male mating success. Individual analyses showed that female visits generally increased with lek size across the majority of taxa surveyed; the meta-analysis indicated that this relationship with lek size was disproportionately positive. The findings from analysing per capita female visits were mixed, with an increase with lek size detected in half of the species, which were, however, widely distributed taxonomically. Taken together, these findings suggest that a female bias for clustered males may be a general process across lekking species. Nevertheless, the substantial variation seen in these relationships implies that other processes are also important. Analyses of per capita copulation success suggested that, more generally, increased per capita mating benefits may be an important selective factor in lek maintenance.

Scalable Preference Aggregation in Social Networks

In social choice theory, preference aggregation refers to computing an aggregate preference over a set of alternatives given individual preferences of all the agents. In real-world scenarios, it may not be feasible to gather preferences from all the agents. Moreover, determining the aggregate preference is computationally intensive. In this paper, we show that the aggregate preference of the agents in a social network can be computed efficiently and with sufficient accuracy using preferences elicited from a small subset of critical nodes in the network. Our methodology uses a model developed based on real-world data obtained using a survey on human subjects, and exploits network structure and homophily of relationships. Our approach guarantees good performance for aggregation rules that satisfy a property which we call expected weak insensitivity. We demonstrate empirically that many practically relevant aggregation rules satisfy this property. We also show that two natural objective functions in this context satisfy certain properties, which makes our methodology attractive for scalable preference aggregation over large scale social networks. We conclude that our approach is superior to random polling while aggregating preferences related to individualistic metrics, whereas random polling is acceptable in the case of social metrics.

Triarylborane-dipyrromethane conjugates bearing dual receptor sites: the synthesis and evaluation of the anion binding site preference

The synthesis and optical properties of four new triarylborane-dipyrromethane (TAB-DPM) conjugates (3a-d) containing dual binding sites (hydrogen bond donor and Lewis acid) have been reported. The new compounds exhibit a selective fluorogenic response towards the F-ion. The NMR titrations show that the anions bind to the TAB-DPM conjugates via the Lewis acidic triarylborane centre in preference to the hydrogen bond donor (dipyrromethane) units.

Tautomeric Preference and Conformation Locking in Fenobam, Thiofenobam, and Their Analogues: The Decisive Role of Hydrogen Bond Hierarchy

The crystal and molecular structures of the potential antidepressant drug fenobam and its derivatives are examined in terms of the preferred form among the two possible tautomeric structures. In this study, chemical derivatization has been utilized as a means to ``experimentally simulate'' the tautomeric preference and conformational variability in fenobam. Eight new derivatives of fenobam have been synthesized, and structural features have been characterized by single-crystal X-ray diffraction and NMR spectroscopy. The specific tautomeric preference found in all of these compounds and their known crystal forms have been construed in terms of the stabilizing intramolecular N-H center dot center dot center dot O and N-H center dot center dot center dot S hydrogen bonding. The hierarchy of intramolecular hydrogen bonds evidenced as the preference of the C-H center dot center dot center dot O hydrogen bond over C-H center dot center dot center dot N and that of the C-H center dot center dot center dot N hydrogen bond over C-H center dot center dot center dot S explains the two distinct conformations adopted by fenobam and thiofenobam derivatives. The relative energy values of different molecular conformations have been calculated and compared.

A Comparative Assessment of Hand Preference in Captive Red Howler Monkeys, Alouatta seniculus and Yellow-Breasted Capuchin Monkeys, Sapajus xanthosternos

There are two major theories that attempt to explain hand preference in non-human primates-the `task complexity' theory and the `postural origins' theory. In the present study, we proposed a third hypothesis to explain the evolutionary origin of hand preference in non-human primates, stating that it could have evolved owing to structural and functional adaptations to feeding, which we refer to as the `niche structure' hypothesis. We attempted to explore this hypothesis by comparing hand preference across species that differ in the feeding ecology and niche structure: red howler monkeys, Alouatta seniculus and yellow-breasted capuchin monkeys, Sapajus xanthosternos. The red howler monkeys used the mouth to obtain food more frequently than the yellow-breasted capuchin monkeys. The red howler monkeys almost never reached for food presented on the opposite side of a wire mesh or inside a portable container, whereas the yellow-breasted capuchin monkeys reached for food presented in all four spatial arrangements (scattered, on the opposite side of a wire mesh, inside a suspended container, and inside a portable container). In contrast to the red howler monkeys that almost never acquired bipedal and clinging posture, the yellow-breasted capuchin monkeys acquired all five body postures (sitting, bipedal, tripedal, clinging, and hanging). Although there was no difference between the proportion of the red howler monkeys and the yellow-breasted capuchin monkeys that preferentially used one hand, the yellow-breasted capuchin monkeys exhibited an overall weaker hand preference than the red howler monkeys. Differences in hand preference diminished with the increasing complexity of the reaching-for-food tasks, i.e., the relatively more complex tasks were perceived as equally complex by both the red howler monkeys and the yellow-breasted capuchin monkeys. These findings suggest that species-specific differences in feeding ecology and niche structure can influence the perception of the complexity of the task and, consequently, hand preference.

First principles study of structural stability and site preference in Co-3 (W,X)

Since the discovery 1] of gamma' precipitate (L1(2) - Co-3 (Al, W)) in the Co-Al-W ternary system, there has been an increased interest in Co-based superalloys. Since these alloys have two phase microstructures (gamma + gamma') similar to Ni-based superalloys 2], they are viable candidates in high temperature applications, particularly in land-based turbines. The role of alloying on stability of the gamma' phase has been an active area of research. In this study, electronic structure calculations were done to probe the effect of alloying in Co3W with L1(2) structure. Compositions of type Co-3(W, X), (where X/Y = Mn, Fe, Ni, Pt, Cr, Al, Si, V, W, Ta, Ti, Nb, Hf, Zr and Mo) were studied. Effect of alloying on equilibrium lattice parameters and ground state energies was used to calculate Vegard's coefficients and site preference related data. The effect of alloying on the stability of the L1(2) structure vis a vis other geometrically close packed ordered structures was also studied for a range of Co3X compounds. Results suggest that the penchant of element for the W sublattice can be predicted by comparing heats of formation of Co3X in different structures.