Improving query focused summarization using look-ahead strategy

Query focused summarization is the task of producing a compressed text of original set of documents based on a query. Documents can be viewed as graph with sentences as nodes and edges can be added based on sentence similarity. Graph based ranking algorithms which use 'Biased random surfer model' like topic-sensitive LexRank have been successfully applied to query focused summarization. In these algorithms, random walk will be biased towards the sentences which contain query relevant words. Specifically, it is assumed that random surfer knows the query relevance score of the sentence to where he jumps. However, neighbourhood information of the sentence to where he jumps is completely ignored. In this paper, we propose look-ahead version of topic-sensitive LexRank. We assume that random surfer not only knows the query relevance of the sentence to where he jumps but he can also look N-step ahead from that sentence to find query relevance scores of future set of sentences. Using this look ahead information, we figure out the sentences which are indirectly related to the query by looking at number of hops to reach a sentence which has query relevant words. Then we make the random walk biased towards even to the indirect query relevant sentences along with the sentences which have query relevant words. Experimental results show 20.2% increase in ROUGE-2 score compared to topic-sensitive LexRank on DUC 2007 data set. Further, our system outperforms best systems in DUC 2006 and results are comparable to state of the art systems.

Mapping the apoptosis inducing domain of an immunomodulatory protein: glycodelin A

Glycodelin A (GdA) is a dimeric glycoprotein synthesized by the human endometrium under progesterone regulation. Based on the high sequence similarity with beta-lactoglobulin, it is placed under the lipocalin superfamily. The protein is one of the local immunomodulators present at the feto-maternal interface which affects both the innate as well as the acquired arms of the immune system, thereby bringing about successful establishment and progression of pregnancy. Our previous studies revealed that the domain responsible for the immunosuppressive activity of glycodelin lies on its protein backbone and the glycans modulate the same. This study attempts to further delineate the apoptosis inducing region of GdA. Our results demonstrate that the stretch of amino acid sequence between Met24 to Leu105 is necessary and sufficient to inhibit proliferation of T cells and induce apoptosis in them. Further, within this region the key residues involved in harboring the activity were shown to be present between Asp52 and Ser65.

Structural insights into the role of Bacillus subtilis YwfH (BacG) in tetrahydrotyrosine synthesis

The synthesis of the dipeptide antibiotic bacilysin involves the sequential action of multiple enzymes in the bac operon. YwfH (also referred to as BacG) catalyzes the stereoselective reduction of dihydro-hydroxyphenylpyruvate (H2HPP) to tetrahydro-hydroxyphenylpyruvate (H4HPP) in this biosynthetic pathway. YwfH is an NADPH-dependent reductase that facilitates the conjugate addition of a hydride at the C4 olefin terminus of H2HPP. Here, the structure of YwfH is described at three conformational steps: the apo form, an apo-like conformation and the NADPH complex. YwfH is structurally similar to other characterized short-chain dehydrogenase/reductases despite having marginal sequence similarity. The structures of YwfH in different conformational states provide a rationale for the ping-pong reaction mechanism. The identification and role of the residues in the catalytic tetrad (Lys113Tyr117Ser155Asn158) in proton transfer were examined by mutational analysis. Together, the structures and biochemical features revealed synchronized conformational changes that facilitate cofactor specificity and catalysis of H4HPP formation en route to tetrahydrotyrosine synthesis.

Obtaining single document summaries using latent dirichlet allocation

In this paper, we present a novel approach that makes use of topic models based on Latent Dirichlet allocation(LDA) for generating single document summaries. Our approach is distinguished from other LDA based approaches in that we identify the summary topics which best describe a given document and only extract sentences from those paragraphs within the document which are highly correlated given the summary topics. This ensures that our summaries always highlight the crux of the document without paying any attention to the grammar and the structure of the documents. Finally, we evaluate our summaries on the DUC 2002 Single document summarization data corpus using ROUGE measures. Our summaries had higher ROUGE values and better semantic similarity with the documents than the DUC summaries.

Mycobacterium tuberculosis RsdA provides a conformational rationale for selective regulation of sigma-factor activity by proteolysis

The relative levels of different sigma factors dictate the expression profile of a bacterium. Extracytoplasmic function sigma factors synchronize the transcriptional profile with environmental conditions. The cellular concentration of free extracytoplasmic function sigma factors is regulated by the localization of this protein in a sigma/anti-sigma complex. Anti-sigma factors are multi-domain proteins with a receptor to sense environmental stimuli and a conserved anti-sigma domain (ASD) that binds a sigma factor. Here we describe the structure of Mycobacterium tuberculosis anti-sigma(D) (RsdA) in complex with the -35 promoter binding domain of sigma(D) (sigma(D)(4)). We note distinct conformational features that enable the release of sigma(D) by the selective proteolysis of the ASD in RsdA. The structural and biochemical features of the sigma(D)/RsdA complex provide a basis to reconcile diverse regulatory mechanisms that govern sigma/anti-sigma interactions despite high overall structural similarity. Multiple regulatory mechanisms embedded in an ASD scaffold thus provide an elegant route to rapidly re-engineer the expression profile of a bacterium in response to an environmental stimulus.

Model-free simulations for compressible mixing layer

Confined supersonic mixing layer is explored through model-free simulations. Both two- and three-dimensional spatio-temporal simulations were carried out employing higher order finite difference scheme as well as finite volume scheme based on open source software (OpenFOAM) to understand the effect of three-dimensionality on the development of mixing layer. It is observed that although the instantaneous structures exhibit three-dimensional features, the average pressure and velocities are predominantly two-dimensional. The computed wall pressures match well with experimental results fairly well, although three-dimensional simulation underpredicts the wall pressure in the downstream direction. The self-similarity of the velocity profiles is obtained within the duct length for all the simulations. Although the mixing layer thicknesses differ among different simulations, their growth rate is nearly the same. Significant differences are observed for species and temperature distribution between two- and three-dimensional calculations, and two-dimensional calculations do not match the experimental observation of smooth variations in species mass fraction profiles as reported in literature. Reynolds stress distribution for three-dimensional calculations show profiles with less peak values compared to two-dimensional calculations; while normal stress anisotropy is higher for three-dimensional case.

A thermodynamic framework for fatigue crack growth in concrete

The phenomenon of fatigue is commonly observed in majority of concrete structures and it is important to mathematically model it in order to predict their remaining life. An energy approach is adopted in this research by using the framework of thermodynamics wherein the dissipative phenomenon is described by a dissipation potential. An analytical expression is derived for the dissipation potential using the concepts of dimensional analysis and self-similarity to describe a fatigue crack propagation model for concrete. This is validated using available experimental results. Through a sensitivity analysis, the hierarchy of importance of different parameters is highlighted.

Deformation behaviour at macro- and nano-length scales: The development of orientation gradients

The deformation behaviour of macrocrystalline and nanocrystalline nickel shows a striking similarity in terms of higher intragranular misorientation and a texture with dominant Brass component on rolling. This is in contrast to microcrystalline nickel, with lower intragranular misorientation and typical Copper type texture. This has been attributed to the free surfaces in macrocrystalline sample and grain boundaries in nanocrystalline sample. Experimental evidence of `Grain Boundary Affected Zone' (GBAZ) showing multi-slip in contrast to limited slip in the grain interiors has been provided. The similarity in evolution of texture and intragranular misorientation is explained on the basis of reduced contribution from the GBAZ at the two extreme length scales.

Prioritisation of micro-catchments based on morphology

Two multicriterion decision-making methods, namely `compromise programming' and the `technique for order preference by similarity to an ideal solution' are employed to prioritise 22 micro-catchments (A1 to A22) of Kherthal catchment, Rajasthan, India and comparative analysis is performed using the compound parameter approach. Seven criteria - drainage density, bifurcation ratio, stream frequency, form factor, elongation ratio, circulatory ratio and texture ratio - are chosen for the evaluation. The entropy method is employed to estimate weights or relative importance of the criterion which ultimately affects the ranking pattern or prioritisation of micro-catchments. Spearman rank correlation coefficients are estimated to measure the extent to which the ranks obtained are correlated. Based on the average ranking approach supported by sensitivity analysis, micro-catchments A6, A10, A3 are preferred (owing to their low ranking) for further improvements with suitable conservation and management practices, and other micro-catchments can be processed accordingly at a later phase on a priority basis. It is concluded that the present approach can be explored for other similar situations with appropriate modifications.

Supplier selection based on supply chain ecosystem, performance and risk criteria

A supply chain ecosystem consists of the elements of the supply chain and the entities that influence the goods, information and financial flows through the supply chain. These influences come through government regulations, human, financial and natural resources, logistics infrastructure and management, etc., and thus affect the supply chain performance. Similarly, all the ecosystem elements also contribute to the risk. The aim of this paper is to identify both performances-based and risk-based decision criteria, which are important and critical to the supply chain. A two step approach using fuzzy AHP and fuzzy technique for order of preference by similarity to ideal solution has been proposed for multi-criteria decision-making and illustrated using a numerical example. The first step does the selection without considering risks and then in the next step suppliers are ranked according to their risk profiles. Later, the two ranks are consolidated into one. In subsequent section, the method is also extended for multi-tier supplier selection. In short, we are presenting a method for the design of a resilient supply chain, in this paper.

Multi-task learning using shared and task specific information

Multi-task learning solves multiple related learning problems simultaneously by sharing some common structure for improved generalization performance of each task. We propose a novel approach to multi-task learning which captures task similarity through a shared basis vector set. The variability across tasks is captured through task specific basis vector set. We use sparse support vector machine (SVM) algorithm to select the basis vector sets for the tasks. The approach results in a sparse model where the prediction is done using very few examples. The effectiveness of our approach is demonstrated through experiments on synthetic and real multi-task datasets.

Multiscaling in Hall-Magnetohydrodynamic Turbulence: Insights from a Shell Model

We show that a shell-model version of the three-dimensional Hall-magnetohydrodynamic (3D Hall-MHD) equations provides a natural theoretical model for investigating the multiscaling behaviors of velocity and magnetic structure functions. We carry out extensive numerical studies of this shell model, obtain the scaling exponents for its structure functions, in both the low-k and high-k power-law ranges of three-dimensional Hall-magnetohydrodynamic, and find that the extended-self-similarity procedure is helpful in extracting the multiscaling nature of structure functions in the high-k regime, which otherwise appears to display simple scaling. Our results shed light on intriguing solar-wind measurements.

The Role of Weak Interactions in the Phase Transition and Distinct Mechanical Behavior of Two Structurally Similar Caffeine Co-crystal Polymorphs Studied by Nanoindentation

Although weak interactions, such as C-H center dot center dot center dot O and pi-stacking, are generally considered to be insignificant, it is their reorganization that holds the key for many a solid-state phenomenon, such as phase transitions, plastic deformation, elastic flexibility, and mechanochromic luminescence in solid-state fluorophores. Despite this, the role of weak interactions in these dynamic phenomena is poorly understood. In this study, we investigate two co-crystal polymorphs of caffeine:4-chloro-3-nitrobenzoic acid, which have close structural similarity (2D layered structures), but surprisingly show distinct mechanical behavior. Form I is brittle, but shows shear-induced phase instability and, upon grinding, converts to Form II, which is soft and plastically shearable. This observation is in contrast to those reported in earlier studies on aspirin, wherein the metastable drug forms are softer and convert to stable and harder forms upon stressing To establish a molecular level understanding, have investigated the two co-crystal polymorphs I and II by single crystal X-ray diffraction, nanoindentation to quantify mechanical properties, and theoretical calculations. The lower hardness (from nanoindentation) and smooth potential surfaces (from theoretical studies) for shearing of layers in Form II allowed us to rationalize the role of stronger intralayer (sp(2))C-H center dot center dot center dot O and nonspecific interlayer pi-stacking interactions in the structure of II. Although the Form I also possesses the same type of interactions, its strength is clearly opposite, that is, weaker intralayer (sp(3))C-H center dot center dot center dot O and specific interlayer pi-stacking interactions. Hence, Form I is harder than Form IL Theoretical calculations and indentation on (111) of Form I suggested the low resistance of this face to mechanical stress; thus, Form I converts to II upon mechanical action. Hence, our approach demonstrates the usefulness of multiple techniques for establishing the role of weak noncovalent interactions in solid-state dynamic phenomena, such as stress induced phase transformation, and hence is important in the context of solid-state pharmaceutical chemistry and crystal engineering.

Does linear separability really matter? Complex visual search is explained by simple search

Visual search in real life involves complex displays with a target among multiple types of distracters, but in the laboratory, it is often tested using simple displays with identical distracters. Can complex search be understood in terms of simple searches? This link may not be straightforward if complex search has emergent properties. One such property is linear separability, whereby search is hard when a target cannot be separated from its distracters using a single linear boundary. However, evidence in favor of linear separability is based on testing stimulus configurations in an external parametric space that need not be related to their true perceptual representation. We therefore set out to assess whether linear separability influences complex search at all. Our null hypothesis was that complex search performance depends only on classical factors such as target-distracter similarity and distracter homogeneity, which we measured using simple searches. Across three experiments involving a variety of artificial and natural objects, differences between linearly separable and nonseparable searches were explained using target-distracter similarity and distracter heterogeneity. Further, simple searches accurately predicted complex search regardless of linear separability (r = 0.91). Our results show that complex search is explained by simple search, refuting the widely held belief that linear separability influences visual search.

Quantifying Vocal Mimicry in the Greater Racket-Tailed Drongo: A Comparison of Automated Methods and Human Assessment

Objective identification and description of mimicked calls is a primary component of any study on avian vocal mimicry but few studies have adopted a quantitative approach. We used spectral feature representations commonly used in human speech analysis in combination with various distance metrics to distinguish between mimicked and non-mimicked calls of the greater racket-tailed drongo, Dicrurus paradiseus and cross-validated the results with human assessment of spectral similarity. We found that the automated method and human subjects performed similarly in terms of the overall number of correct matches of mimicked calls to putative model calls. However, the two methods also misclassified different subsets of calls and we achieved a maximum accuracy of ninety five per cent only when we combined the results of both the methods. This study is the first to use Mel-frequency Cepstral Coefficients and Relative Spectral Amplitude - filtered Linear Predictive Coding coefficients to quantify vocal mimicry. Our findings also suggest that in spite of several advances in automated methods of song analysis, corresponding cross-validation by humans remains essential.