Dependency Tree Research Articles

Graph Convolutional Networks With Argument-Aware Pooling for Event Detection

The current neural network models for event detection have only considered the sequential representation of sentences. Syntactic representations have not been explored in this area although they provide an effective mechanism to directly link words to their informative context for event detection in the sentences. In this work, we investigate a convolutional neural network based on dependency trees to perform event detection. We propose a novel pooling method that relies on entity mentions to aggregate the convolution vectors. The extensive experiments demonstrate the benefits of the dependency-based convolutional neural networks and the entity mention-based pooling method for event detection. We achieve the state-of-the-art performance on widely used datasets with both perfect and predicted entity mentions.

LPTK: a linguistic pattern-aware dependency tree kernel approach for the BioCreative VI CHEMPROT task.

Identifying the interactions between chemical compounds and genes from biomedical literatures is one of the frequently discussed topics of text mining in the life science field. In this paper, we describe Linguistic Pattern-Aware Dependency Tree Kernel, a linguistic interaction pattern learning method developed for CHEMPROT task–BioCreative VI, to capture chemical–protein interaction (CPI) patterns within biomedical literatures. We also introduce a framework to integrate these linguistic patterns with smooth partial tree kernel to extract the CPIs. This new method of feature representation models aspects of linguistic probability in geometric representation, which not only optimizes the sufficiency of feature dimension for classification, but also defines features as interpretable contexts rather than long vectors of numbers. In order to test the robustness and efficiency of our system in identifying different kinds of biological interactions, we evaluated our framework on three separate data sets, i.e. CHEMPROT corpus, Chemical–Disease Relation corpus and Protein–Protein Interaction corpus. Corresponding experiment results demonstrate that our method is effective and outperforms several compared systems for each data set.

A GRAPH-THEORETIC ANALYSIS OF THE SEMANTIC PARADOXES

AbstractWe introduce a framework for a graph-theoretic analysis of the semantic paradoxes. Similar frameworks have been recently developed for infinitary propositional languages by Cook [5, 6] and Rabern, Rabern, and Macauley [16]. Our focus, however, will be on the language of first-order arithmetic augmented with a primitive truth predicate. Using Leitgeb’s [14] notion of semantic dependence, we assign reference graphs (rfgs) to the sentences of this language and define a notion of paradoxicality in terms of acceptable decorations of rfgs with truth values. It is shown that this notion of paradoxicality coincides with that of Kripke [13]. In order to track down the structural components of an rfg that are responsible for paradoxicality, we show that any decoration can be obtained in a three-stage process: first, the rfg is unfolded into a tree, second, the tree is decorated with truth values (yielding a dependence tree in the sense of Yablo [21]), and third, the decorated tree is re-collapsed onto the rfg. We show that paradoxicality enters the picture only at stage three. Due to this we can isolate two basic patterns necessary for paradoxicality. Moreover, we conjecture a solution to the characterization problem for dangerous rfgs that amounts to the claim that basically the Liar- and the Yablo graph are the only paradoxical rfgs. Furthermore, we develop signed rfgs that allow us to distinguish between ‘positive’ and ‘negative’ reference and obtain more fine-grained versions of our results for unsigned rfgs.

Linguistic $$\leftrightarrow $$ ↔ Rational Agents’ Semantics

We define and prove a formal semantics divided into two complementary interacting components: the strictly linguistic (i.e. linguistically marked) semantics, we call linguistic agent (LA), and the strictly logical and referential semantics, we call rational agent (RA). This Linguistic $$\leftrightarrow $$ Rational Agents’ Semantics (LRA semantics) applies to Deep Dependency trees (DD-trees) or more generally, to discourses, i.e. sequences of DD-trees, and interprets them by functional structures we call Meaning Representation Structures (MRS), similar to the DRT, but interpreted very differently. LRA semantics incrementally interprets the discourses by minimal finite models, called proto-models, in a monotonic logic of the LA and checks the proto-models with respect to the classical models of the RA. The proto-model is considered as the linguistic sense of the discourse. We define in full detail the LA which, as we believe, must be universal. On the other hand, we don’t propose a particular RA. We only define the scheme of interaction between the two agents and the stimuli of the RA used by the LA. After all, every discourse has in LRA semantics the single meaning and the single sense for every Rational Agent used to interact with the Linguistic Agent.

AN EFFICIENT APPROACH OF REGRESSION TESTING USING HIERARCHICAL DECOMPOSITION SLICING

The purpose of regression testing is to ensure that bug fixes and new functionality introduced in a new version of a software do not adversely affect the correct functionality inherited from the previous version. I have proposed a new slicing method to decompose a Java program into affected packages, classes, methods and statements identified with respect to the modification made in the program.Because java program has hierarchical characteristics(Package/Class/Method), this decomposition is based on this characteristic. I also have proposed an intermediate representation for Java programs that shows Package Level, Class Level, and Method Level Dependency. I have named this intermediate representation as Program Dependency Tree(PDT). This PDT is used to identify the program parts that are possibly affected by the change made to the program. With the help of PDT, we can obtain program slices. The packages, classes, methods, and statements thus affected are identified by traversing the Program Dependency Tree. The proposed approach maps the decomposed slice (comprising of the affected program parts) with the coverage information of the existing test suite to select the appropriate test cases for regression testing.

Applying intelligent methods in detecting maritime smuggling

Taiwan is an island state characterized by well-developed maritime traffic and sea transportation. Smuggling by sea increasingly threatens the country’s social and internal security. The Coast Guard Administration is the main actor intercepting smuggling. This study applies classification trees and Bayesian network algorithms to create dependent and independent classification tree models and Bayesian network models for improving the smuggling vessel recognition rates. We focus on relationships of dependence among smuggled goods (for example, fishery goods, firearms and drugs) and variables (such as information on the home port, the age of boat owners, as well as the arrival and departure dates of vessels). The paper presents a selection method for vessels which could be applicable in a number of other maritime instances, such as, for instance, Port State Control inspection. Our findings help in improving recognition accuracy and reduce the potential harm to social and national security by facilitating the identification of vessels with the largest smuggling value.

Tradeoff between execution speedup and reliability for compute-intensive code offloading in mobile device cloud

With the advent of different mobile computing technologies, mobile devices have opened up a plethora of computational infrastructure to provide improved performance for compute-intensive applications to the end users. Mobile Device Cloud (MDC) technology brings the code offloading mechanism from distant cloud to neighbor mobile devices. However, the major challenges of code offloading in MDC systems include maximization of computation speedup and reliability; unfortunately, these two performance parameters often oppose each other. In this paper, an optimization framework, namely TESAR, has been devised to tradeoff between application execution speedup and reliability while maintaining device energy within a predefined range. We also provide an algorithm for developing a dependency tree among the modules of an application so as to allow higher number of parallel executions, wherever and whenever it is possible. The emulation results of the proposed algorithm outperform the relevant state-of-the-art works in terms of application completion time, communication latency and rescheduling overhead.

A CDT-Styled End-to-End Chinese Discourse Parser

Discourse parsing is a challenging task and plays a critical role in discourse analysis. Since the release of the Rhetorical Structure Theory Discourse Treebank and the Penn Discourse Treebank, the research on English discourse parsing has attracted increasing attention and achieved considerable success in recent years. At the same time, some preliminary research on certain subtasks about discourse parsing for other languages, such as Chinese, has been conducted. In this article, we present an end-to-end Chinese discourse parser with the Connective-Driven Dependency Tree scheme, which consists of multiple components in a pipeline architecture, such as the elementary discourse unit (EDU) detector, discourse relation recognizer, discourse parse tree generator, and attribution labeler. In particular, the attribution labeler determines two attributions (i.e., sense and centering) for every nonterminal node (i.e., discourse relation) in the discourse parse trees. Systematically, our parser detects all EDUs in a free text, generates the discourse parse tree in a bottom-up way, and determines the sense and centering attributions for all nonterminal nodes by traversing the discourse parse tree. Comprehensive evaluation on the Connective-Driven Dependency Treebank corpus from both component-wise and error-cascading perspectives is conducted to illustrate how each component performs in isolation, and how the pipeline performs with error propagation. Finally, it shows that our end-to-end Chinese discourse parser achieves an overall F1 score of 20% with full automation.

Multivariate extreme value copulas with factor and tree dependence structures

Parsimonious extreme value copula models with O(d) parameters for d observed variables of extrema are presented. These models utilize the dependence characteristics, including factor and tree structures, assumed on the underlying variables that give rise to the data of extremes. For factor structures, a class of parametric models is obtained by taking the extreme value limit of factor copulas with non-zero tail dependence. An alternative model suitable for both factor and tree structures imposes constraints on the parametric Husler-Reiss copula to get representations in terms of O(d) other parameters. Dependence properties are discussed. As the full density is often intractable, the method of composite (pairwise) likelihood is used for model inference. Procedures to improve the stability of bivariate density evaluation are also developed. The proposed models are applied to two data examples — one for annual extreme river flows and one for bimonthly extremes of daily stock returns.

Minimum spanning tree filtering of correlations for varying time scales and size of fluctuations.

Based on a recently proposed q-dependent detrended cross-correlation coefficient, ρ_{q} [J. Kwapień, P. Oświęcimka, and S. Drożdż, Phys. Rev. E 92, 052815 (2015)PLEEE81539-375510.1103/PhysRevE.92.052815], we generalize the concept of the minimum spanning tree (MST) by introducing a family of q-dependent minimum spanning trees (qMSTs) that are selective to cross-correlations between different fluctuation amplitudes and different time scales of multivariate data. They inherit this ability directly from the coefficients ρ_{q}, which are processed here to construct a distance matrix being the input to the MST-constructing Kruskal's algorithm. The conventional MST with detrending corresponds in this context to q=2. In order to illustrate their performance, we apply the qMSTs to sample empirical data from the American stock market and discuss the results. We show that the qMST graphs can complement ρ_{q} in disentangling "hidden" correlations that cannot be observed in the MST graphs based on ρ_{DCCA}, and therefore, they can be useful in many areas where the multivariate cross-correlations are of interest. As an example, we apply this method to empirical data from the stock market and show that by constructing the qMSTs for a spectrum of q values we obtain more information about the correlation structure of the data than by using q=2 only. More specifically, we show that two sets of signals that differ from each other statistically can give comparable trees for q=2, while only by using the trees for q≠2 do we become able to distinguish between these sets. We also show that a family of qMSTs for a range of q expresses the diversity of correlations in a manner resembling the multifractal analysis, where one computes a spectrum of the generalized fractal dimensions, the generalized Hurst exponents, or the multifractal singularity spectra: the more diverse the correlations are, the more variable the tree topology is for different q's. As regards the correlation structure of the stock market, our analysis exhibits that the stocks belonging to the same or similar industrial sectors are correlated via the fluctuations of moderate amplitudes, while the largest fluctuations often happen to synchronize in those stocks that do not necessarily belong to the same industry.

Assessing water use by Prosopis invasions and Vachellia karroo trees: Implications for groundwater recovery following alien plant removal in an arid catchment in South Africa

Prosopis spp., (or Mesquite), is a desert adapted woody weed which has invaded large parts of arid and semi-arid regions. It has a characteristic deep tap root, and forms dense impenetrable thickets along river banks, flood plains, and seasonal water courses. Several studies have quantified the water use by Prosopis, but there is limited information on how this compares with that of co-occurring groundwater dependent indigenous tree species in the arid and semi-arid tropics. Consequently the effects of removing Prosopis on groundwater, where they co-occur with indigenous trees, are not known. This study quantified the incremental water use by Prosopis invasions compared with the indigenous deep rooted Vachellia karroo (V. karroo) trees that normally replace Prosopis once it has been cleared. The study was done in an arid catchment in the Northern Cape Province of South Africa. Stem sap flow dynamics, groundwater levels, and weather data were collected hourly over 15months in the invaded catchment. Data collection continued for a further 11months after Prosopis had been cleared to monitor the recovery of the ecosystem from the invasions. There was no significant difference (P=0.116) in the sap flux density of the two species although Prosopis had a much thinner sapwood depth than V. karroo due to its ring porous wood anatomy. However, at the stand scale Prosopis transpired more than five times more water (∼544mmy−1) than V. karroo (∼91mmy−1) since there were six times more Prosopis than the indigenous species. The removal of Prosopis slowed down the rate of decline of the water table from a pre-clearing peak of ∼8.9–5.0mmd−1 after alien plant removal. Tree transpiration was accurately predicted by a simple model that used aquifer properties and hourly changes in water table depths with a RMSE of ±0.22mmd−1 before and ±0.32mmd−1 after clearing Prosopis.

SNP by SNP by environment interaction network of alcoholism

BackgroundAlcoholism has a strong genetic component. Twin studies have demonstrated the heritability of a large proportion of phenotypic variance of alcoholism ranging from 50–80%. The search for genetic variants associated with this complex behavior has epitomized sequence-based studies for nearly a decade. The limited success of genome-wide association studies (GWAS), possibly precipitated by the polygenic nature of complex traits and behaviors, however, has demonstrated the need for novel, multivariate models capable of quantitatively capturing interactions between a host of genetic variants and their association with non-genetic factors. In this regard, capturing the network of SNP by SNP or SNP by environment interactions has recently gained much interest.ResultsHere, we assessed 3,776 individuals to construct a network capable of detecting and quantifying the interactions within and between plausible genetic and environmental factors of alcoholism. In this regard, we propose the use of first-order dependence tree of maximum weight as a potential statistical learning technique to delineate the pattern of dependencies underpinning such a complex trait. Using a predictive based analysis, we further rank the genes, demographic factors, biological pathways, and the interactions represented by our SNP times SNP times E network. The proposed framework is quite general and can be potentially applied to the study of other complex traits.

Efficient Dependency-Guided Named Entity Recognition

Named entity recognition (NER), which focuses on the extraction of semantically meaningful named entities and their semantic classes from text, serves as an indispensable component for several down-stream natural language processing (NLP) tasks such as relation extraction and event extraction. Dependency trees, on the other hand, also convey crucial semantic-level information. It has been shown previously that such information can be used to improve the performance of NER. In this work, we investigate on how to better utilize the structured information conveyed by dependency trees to improve the performance of NER. Specifically, unlike existing approaches which only exploit dependency information for designing local features, we show that certain global structured information of the dependency trees can be exploited when building NER models where such information can provide guided learning and inference. Through extensive experiments, we show that our proposed novel dependency-guided NER model performs competitively with models based on conventional semi-Markov conditional random fields, while requiring significantly less running time.

A Deep Learning Approach for Arabic Caption Generation Using Roots-Words

Automatic caption generation is a key research field in the machine learning community. However, most of the current research is performed on English caption generation ignoring other languages like Arabic and Persian. In this paper, we propose a novel technique leveraging the heavy influence of root words in Arabic to automatically generate captions in Arabic. Fragments of the images are associated with root words and deep belief network pre-trained using Restricted Boltzmann Machines are used to extract words associated with image. Finally, dependency tree relations are used to generate sentence-captions by using the dependency on root words. Our approach is robust and attains BLEU-1 score of 34.8.

Hindi CCGbank: A CCG treebank from the Hindi dependency treebank

In this paper, we present an approach for automatically creating a combinatory categorial grammar (CCG) treebank from a dependency treebank for the subject–object–verb language Hindi. Rather than a direct conversion from dependency trees to CCG trees, we propose a two stage approach: a language independent generic algorithm first extracts a CCG lexicon from the dependency treebank. An exhaustive CCG parser then creates a treebank of CCG derivations. We also discuss special cases of this generic algorithm to handle linguistic phenomena specific to Hindi. In doing so we extract different constructions with long-range dependencies like coordinate constructions and non-projective dependencies resulting from constructions like relative clauses, noun elaboration and verbal modifiers.

Nonoptimality of the Maximum-Weight Dependence Tree in Classification

Half a century ago, Chow and Liu proved that the distribution of the first-order dependence tree that optimally approximates an arbitrary joint distribution in terms of Kullback–Leibler cross entropy is the distribution of the maximum-weight dependence tree (MWDT). In a $p$ -dimensional space, this is a tree with $p-1$ branches between pairs of variables with each branch having a weight equal to the mutual information between variables at both ends. Today, MWDTs have applications that are more important in classification than only being used to model first-order dependence structure among variables. Nevertheless, whether the MWDT is the optimal first-order tree in classification has been left unexplored so far. In this letter, we study the optimality of the MWDT structure from the stand-point of classification.

Constrained Partial Parsing Based Dependency Tree Projection for Tree-to-Tree Machine Translation

Constrained Partial Parsing Based Dependency Tree Projection for Tree-to-Tree Machine Translation

Performance of Combined Models in Discrete Binary Classification

Abstract. Diverse Discrete Discriminant Analysis (DDA) models perform differently in different samples. This fact has encouraged research in combined models which seems particularly promising when the a priori classes are not well separated or when small or moderate sized samples are considered, which often occurs in practice. In this study, we evaluate the performance of a convex combination of two DDA models: the First-Order Independence Model (FOIM) and the Dependence Trees Model (DTM). We use simulated data sets with two classes and consider diverse data complexity factors which may influence performance of the combined model – the separation of classes, balance, and number of missing states, as well as sample size and also the number of parameters to be estimated in DDA. We resort to cross-validation to evaluate the precision of classification. The results obtained illustrate the advantage of the proposed combination when compared with FOIM and DTM: it yields the best results, especially when very small samples are considered. The experimental study also provides a ranking of the data complexity factors, according to their relative impact on classification performance, by means of a regression model. It leads to the conclusion that the separation of classes is the most influential factor in classification performance. The ratio between the number of degrees of freedom and sample size, along with the proportion of missing states in the minority class, also has significant impact on classification performance. An additional gain of this study, also deriving from the estimated regression model, is the ability to successfully predict the precision of classification in a real data set based on the data complexity factors.

Constrained Partial Parsing Based Dependency Tree Projection for Tree-to-Tree Machine Translation

Constrained Partial Parsing Based Dependency Tree Projection for Tree-to-Tree Machine Translation

A New Resilience Rating System for Countries and States

This research presents a quantitative method to assess resilience at the state level. The approach introduced in this work is an evolution of the risk assessment concept. Risk is mainly a function of vulnerability, hazard, and exposure; on the other hand, resilience focuses more on the internal characteristics of a system rather than its vulnerability. To tackle this difference, a new formulation has been introduced for the evaluation of resilience. In this formulation, resilience is a function of hazard, exposure, and intrinsic resilience. Generally, intrinsic resilience deals with the internal characteristics of a system, and it differs from the traditional resilience index that takes into account external factors in its assessment, such as the disaster intensity and the level of exposure. The paper also provides a method to compute the intrinsic resilience of countries. This method is based on the data provided by Hyogo Framework for Action (HFA), which is a work developed by the United Nations (UN). HFA evaluates the inherent resilience of countries based on a number of equally weighted indicators. However, further analysis has shown that the contribution made by each of those indicators toward the intrinsic resilience is different. This discrepancy has necessitated weighting the indicators based on their individual contribution towards the intrinsic resilience. To do that, we introduce the Dependence Tree Analysis (DTA). DTA is a method that determines the correlation between a component and its sub-components (i.e., between intrinsic resilience and its indicators), enabling us to orderly allocate new weights to the indicators to obtain a more representative output for the intrinsic resilience. Finally, a case study composed of 37 states has been conducted in order to illustrate the methodology in all details. Both intrinsic resilience and resilience indexes for each of the states were assessed. This was followed by a comparative analysis in order to test the applicability of the methodology, and the results were in line with the predictions.