Dependency Tree Research Articles

L’amas verbal au cœur d’une modélisation topologique du français

This article aims at two objectives : We want to show that principal word order phenomena can be described in a topological approach to linearization, i.e. a linking of an (unordered) dependency tree and a ordered structure, the topological constituent tree. Moreover, we want to put forward the importance and naturalness of a particular constituent in the modeling of French word order, the verb cluster (amas verbal) : It is composed of one or more verbs, of necessary functional words (prepositions and complementizers), and of very constraint lexical elements other than the verbs as for example clitics and certain adverbs. The work is based on a rigorous formalization of the dependency-topology link in Gerdes & Kahane 2001. The verb cluster in the topological structure remains to be shown to be closely related to a prosodic constituent.

The Costs of Ignoring High-Order Correlations in Populations of Model Neurons

Investigators debate the extent to which neural populations use pairwise and higher-order statistical dependencies among neural responses to represent information about a visual stimulus. To study this issue, three statistical decoders were used to extract the information in the responses of model neurons about the binocular disparities present in simulated pairs of left-eye and right-eye images: (1) the full joint probability decoder considered all possible statistical relations among neural responses as potentially important; (2) the dependence tree decoder also considered all possible relations as potentially important, but it approximated high-order statistical correlations using a computationally tractable procedure; and (3) the independent response decoder, which assumed that neural responses are statistically independent, meaning that all correlations should be zero and thus can be ignored. Simulation results indicate that high-order correlations among model neuron responses contain significant information about binocular disparities and that the amount of this high-order information increases rapidly as a function of neural population size. Furthermore, the results highlight the potential importance of the dependence tree decoder to neuroscientists as a powerful but still practical way of approximating high-order correlations among neural responses.

The Costs of Ignoring High-Order Correlations in Populations of Model Neurons

Investigators debate the extent to which neural populations use pair-wise and higher-order statistical dependencies among neural responses to represent information about a visual stimulus. To study this issue, three statistical decoders were used to extract the information in the responses of model neurons about the binocular disparities present in simulated pairs of left-eye and right-eye images: (1) the full joint probability decoder considered all possible statistical relations among neural responses as potentially important; (2) the dependence tree decoder also considered all possible relations as potentially important, but it approximated high-order statistical correlations using a computationally tractable procedure; and (3) the independent response decoder, which assumed that neural responses are statistically independent, meaning that all correlations should be zero and thus can be ignored. Simulation results indicate that high-order correlations among model neuron responses contain significant information about binocular disparities and that the amount of this high-order information increases rapidly as a function of neural population size. Furthermore, the results highlight the potential importance of the dependence tree decoder to neuroscientists as a powerful but still practical way of approximating high-order correlations among neural responses.

Design and Evaluation of a Low-Latency Checkpointing Scheme for Mobile Computing Systems

Fault-tolerant mobile computing systems have different requirements and restrictions, not taken into account by conventional distributed systems. This paper presents a coordinated checkpointing scheme which reduces the delay involved in a global checkpointing process for mobile systems. A piggyback technique is used to track and record the checkpoint dependency information among processes during normal message transmission. During checkpointing, a concurrent checkpointing technique is designed to use the pre-recorded process dependency information to minimize process blocking time by sending checkpoint requests to dependent processes at once, hence saving the time to trace the dependency tree. We show that our checkpoint algorithm forces a minimum number of processes to take checkpoints, which is an important property for checkpointing mobile applications. Via probability-based analysis, we show that our scheme can significantly reduce the latency associated with checkpoint request propagation, compared with traditional coordinated checkpointing approaches. Experimental results indicate that we have <2% overhead in transmitting piggybacked information during normal runtime. However, we can achieve up to a 60% reduction in checkpoint latency time.

Dependency trees in sub-linear time and bounded memory

We focus on the problem of efficient learning of dependency trees. Once grown, they can be used as a special case of a Bayesian network, for PDF approximation, and for many other uses. Given the data, a well-known algorithm can fit an optimal tree in time that is quadratic in the number of attributes and linear in the number of records. We show how to modify it to exploit partial knowledge about edge weights. Experimental results show running time that is near-constant in the number of records, without significant loss in accuracy of the generated trees.

APPROXIMATING DEPENDENCY GRAMMARS THROUGH INTERSECTION OF STAR-FREE REGULAR LANGUAGES

The paper formulates the Hays and Gaifman dependency grammar (HGDG) in terms of constraints on a string based encoding of dependency trees and develops an approach to obtain a regular approximation for these grammars. Our encoding of dependency trees uses brackets in a novel fashion: pairs of brackets indicate dependencies between pairs of positions rather than boundaries of phrases. This leads to several advantages: (i) HGDG rules over the balanced bracketing can be expressed using regular languages. (ii) A new homomorphic representation for context-free languages is obtained. (iii) A star-free regular approximation for the original projective dependency grammar is obtained by limiting the number of stacked dependencies. (iv) By relaxing certain constraints, the encoding can be extended to non-projective dependency trees and graphs, (v) strong generative power of HGDGs can now be characterized through sets of bracketed strings.

Matching Abnormality in Hybrid Machine Translation

Recent MT paradigms, such as example-based MT, pattern-based MT, and hybrid MT contribute to overcoming the shortcoming of rule-based MT by improving the quality of translations. However, a problem with erroneous selection of examples or patterns also exists in these types of MTs. In this paper, we discover a matching abnormality, which we term an Over-Specification Problem that occurs when an input sentence corresponding to the pattern of general sense is actually matched to the patterns of exceptional sense, and thus is incorrectly translated. We formally define the problem and propose a solution based on the similarity of dependency trees and lexical association ratio to detect whether the input tree is dragged into the exceptional range of example trees or not. Finally, promising experimental results are provided. We take a hybrid MT as a testing MT and Chinese-to-Korean as a sample language pair.

LinGO Redwoods

The LinGO Redwoods initiative is a seed activity in the design and development of a new type of treebank. A treebank is a (typically hand-built) collection of natural language utterances and associated linguistic analyses; typical treebanks—as for example the widely recognized Penn Treebank (Marcus, Santorini, & Marcinkiewicz, 1993), the Prague Dependency Treebank (Hajic, 1998), or the German TiGer Corpus (Skut, Krenn, Brants, & Uszkoreit, 1997)—assign syntactic phrase structure or tectogrammatical dependency trees over sentences taken from a naturallyoccuring source, often newspaper text. Applications of existing treebanks fall into two broad categories: (i) use of an annotated corpus in empirical linguistics as a source of structured language data and distributional patterns and (ii) use of the treebank for the acquisition (e.g. using stochastic or machine learning approaches) and evaluation of parsing systems. While several mediumto large-scale treebanks exist for English (and some for other major languages), all pre-existing publicly available resources exhibit the following limitations: (i) the depth of linguistic information recorded in these treebanks is comparatively shallow, (ii) the design and format of linguistic representation in the treebank hard-wires a small, predefined range of ways in which information can be extracted from the treebank, and (iii) representations in existing treebanks are static and over the (often yearor decade-long) evolution of a large-scale treebank tend to fall behind theoretical advances in formal linguistics and grammatical representation. LinGO Redwoods aims at the development of a novel treebanking methodology, (i) rich in nature and dynamic in both (ii) the ways linguistic data can be retrieved from the treebank in varying granularity and (iii) the constant evolution and regular updating of the treebank itself, synchronized to the development of ideas in syntactic theory. Starting in October 2001, the project is aiming to build the foundations for this new type of treebank, develop a basic set of tools required for treebank construction and maintenance, and construct an initial set of 10,000 annotated trees to be distributed together with the tools under an open-source license. Building a largescale treebank, disseminating it, and positioning the corpus as a widely-accepted resource is a multi-year effort; the results of this seeding activity will serve as a proof of concept for the novel approach that is expected to enable the LinGO group at CSLI both to disseminate the approach to the wider academic and industrial audience and to secure appropriate funding for the realization and exploitation of a larger treebank. The purpose of publication at this early stage is three-fold: (i) to encourage feedback on the Redwoods approach from a broader academic audience, (ii) to facilitate exchange with related work at other sites, and (iii) to invite additional collaborators to contribute to the construction of the Redwoods treebank or start its exploitation as early-access versions become available. This paper is an updated version of an earlier project report published by Oepen, Callahan, Flickinger, and Manning (2002); changes over that version include more recent numbers on the current Redwoods development status, inclusion of an example of discriminator-based disambiguation, and minor adaptations and corrections in various parts of the discussion.

Lexicalized Syntax and Topology for Non-projective Dependency Grammar

We propose a lexicalized formulation of dependency grammar that addresses both immediate dependence and linear precedence. Our approach distinguishes two orthogonal, yet mutually constraining dependency trees: an ID tree of syntactic dependencies and a LP tree of topological dependencies. The ID tree is non-ordered, non-projective and its edges are labeled by grammatical functions. The LP tree is ordered and projective and expresses licensed linearizations; its edges are labeled by topological fields. The LP tree can be regarded as deriving from the ID tree through a process of emancipation controlled by lexicalized constraints and principles. In the present article, we formalize valid ID/LP analyses and show how they can be characterized as the solutions of a constraint satisfaction problem. The latter can be solved by constraint programming and forms the basis of our implementation.

3-D Lossless Volumetric Medical Image Compression Using 3-D Integer Wavelet Transform and Lifting Steps

This paper focuses on lossless medical image compression methods for medical images that operate on three-dimensional(3-D) irreversible integer wavelet transform. We offer an application of the Set Partitioning in Hierarchical Trees(SPIHT) algorithm to medical images, using a 3-D wavelet decomposition and a 3-D spatial dependence tree. The wavelet decomposition is accomplished with integer wavelet filters implemented with the lifting method, where careful scaling() and truncations keep the integer precision and the transform unitary. We have tested our encoder on volumetric medical images using different integer filters and different coding unit sizes. The coding unit sizes of 16 slices save considerable dynamic memory(RAM) and coding delay from full sequence coding units used in previous works. Results show that, even with these small coding units, our algorithm with certain filters performs as well and better in lossless coding than previous coding systems using 3-D integer wavelet transforms on volumetric medical images. (Journal of Korean Society of Medical Informatics 10-1, 35-42, 2004)

Comments on "Fast approximation of Kullback-Leibler distance for dependence trees and hidden Markov models"

Comments on "Fast approximation of Kullback-Leibler distance for dependence trees and hidden Markov models"

Fast approximation of Kullback-Leibler distance for dependence trees and hidden Markov models

We present a fast algorithm to approximate the Kullback-Leibler distance (KLD) between two dependence tree models. The algorithm uses the "upward" (or "forward") procedure to compute an upper bound for the KLD. For hidden Markov models, this algorithm is reduced to a simple expression. Numerical experiments show that for a similar accuracy, the proposed algorithm offers a saving of hundreds of times in computational complexity compared to the commonly used Monte Carlo method. This makes the proposed algorithm important for real-time applications, such as image retrieval.

Recent Results in Cophylogeny Mapping

Virtually every problem in biology benefits from consideration within an evolutionary frame work. Parasitism has been part of life ever since one organism was able to provide an environment for another: questions of parasitology naturally lend themselves to consideration of the shared ancient history of parasites and hosts. The derivation of that shared history is therefore an area of great interest to theoreticians and practitioners alike. The most intuitive approach to this is by cophylogeny mapping. Mathematically the problem is that of optimally mapping the dependent tree into the independent one, e.g., parasite into host or gene tree into organismal phylogeny. This article describes some of the recent advances in cophylogenetic simulation, significance testing, and theoretical properties of maps. In simulation the author shows that the number of ways of mapping the parasite phylogeny into that of the hosts does indeed grow exponentially quickly in most cases and shows no close correlation with the similarity between the phylogenies, and that under a simple coevolutionary model, the range of behaviours of simulated parasite phylogenies is extremely broad and would appear to confound efforts to infer model parameters from observed cases. In the area of significance testing the author demonstrates that the maximal number of inferred codivergence events is not necessarily the best statistic for measuring cophylogenetic agreement, and that significance testing by randomisation which does not alter the parasite tree substantially biases results, and provides a new test to determine whether phylogenetic similarity is consistent with preferential host switching.

Topics in decision tree based speech synthesis

Most modern speech synthesis systems using context dependent decision trees in their acoustic synthesis modules are unit selection style concatenative speech synthesis systems using the trees essentially as a form of pruning during their segment search. The IBM Trainable Speech Synthesis System is one such system. This paper begins by discussing the advantages and disadvantages of the decision tree and non-decision tree approaches to unit selection synthesis. It goes on to present the results of formal listening tests conducted on the IBM system to investigate a number of different topics pertinent to decision tree based systems. These include the use of extended context features during clustering, the effect of using trees with different numbers of leaves and different numbers of segments per leaf, and the performance of several different off-line segment preselection algorithms.

Trainable approaches to surface natural language generation and their application to conversational dialog systems

We study how decisions for word ordering and word choice in surface natural language generation can be automatically learned from annotated data. We examine four trainable systems for surface natural language generation in the air travel domain, called NLG[1–4]. NLG1 is a lookup table which stores the most frequent phrase to express a concept, and is intended as a baseline system for comparison purposes. NLG2 and NLG3 attempt to find the highest probability word sequence with respect to a maximum entropy probability model. They differ in that NLG2 predicts words left-to-right, while NLG3 predicts words in dependency tree order. NLG4 requires a dependency-style grammar of phrase fragments and conditions on their use, and attempts to find the highest probability word sequence that is consistent with the rules and conditions of the grammar. NLG4 has been implemented in a dialog strategy for a prototype air travel conversational system, in which word order is dynamically modified to emphasize certain aspects of the run-time dialog state.

Suppletion and dependency in inflectional morphology

This paper presents a general approach to verbal inflection with special emphasis on suppletion phenomena. The paper focuses on French, but the approach is general enough to apply to a wide variety of languages.&#x0D; In the first part of the paper, we show that suppletion is not erratic: suppletive forms tend to always appear in groups, in definite areas of verbal paradigms. Our analysis is based on the observation of a number of dependency relations between inflectional forms of verbs (somewhat similar to rules of referral (Zwicky 1985, Stump 1993)). We define for each language a stem dependency tree based on these observations, which allows one to predict the whole paradigm of every verb in the language on the basis of a minimal number of idiosyncratic stems. We use the tree to minimize the quantity of redundant phonological information that has to be listed in the lexicon for a given lexeme, assuming that an optimal analysis of inflection should be able to derive all and only intuitively predictable inflectional forms from a single representation.&#x0D; The second part of the paper attempts to integrate the analysis in an HPSG hierarchical lexicon. Morphological dependency relations are represented directly by mentioning a lexical sign in another sign's lexical entry. The approach to suppletion proposed in the first part is made explicit using a combination of online type construction and default constraints on the phonology of dependent signs.

Multisource data classification with dependence trees

In order to apply a statistical approach to the classification of multisource remote-sensing data, one of the main problems to face lies in the estimation of probability distribution functions. This problem arises out of the difficulty of defining a common statistical model for such heterogeneous data. A possible solution is to adopt nonparametric approaches, which rely on the availability of training samples without any assumption about the related statistical distributions. The purpose of this paper is to investigate the suitability of the concept of dependence trees for the integration of multisource information through estimation of probability distributions. First, this concept, introduced by Chow and Liu (1968), is used to provide an approximation of a probability distribution defined in an N-dimensional space by a product of N-1 probability distributions defined in two-dimensional (2-D) spaces; this approximation corresponds, in terms of graph theoretical interpretation, to a tree of dependence. For each land cover class, a dependence tree is generated by minimizing an appropriate closeness measure. Then, a nonparametric estimation of the second-order probability distributions is carried out through the Parzen window approach, based on the implementation of 2-D Gaussian kernels. In this way, it is possible to reduce the complexity of the estimation, while capturing a significant part of the interdependence among variables. A comparison with other multisource data fusion methods, namely, the multilayer perceptron (MLP) method, the k-nearest neighbor (k-NN) method, and a Bayesian hierarchical classifier (BHC), is made. Experimental results obtained on multisensor [airborne thematic mapper (ATM) and synthetic aperture radar (SAR)] and multisource (experimental synthetic aperture radar (E-SAR) and a textural feature) data sets show that the proposed fusion method based on dependence trees is able to provide a classification accuracy similar to those of the other methods considered, but with the advantage of a reduced computational load.

Lower Bounds for Dynamic Algebraic Problems

We consider dynamic evaluation of algebraic functions (matrix multiplication, determinant, convolution, Fourier transform, etc.) in the model of Reif and Tate; i.e., if f(x1,…, xn)=(y1, …, ym) is an algebraic problem, we consider serving online requests of the form “change input xi to value v” or “what is the value of output yi?” We present techniques for showing lower bounds on the worst case time complexity per operation for such problems. The first gives lower bounds in a wide range of rather powerful models (for instance, history dependent algebraic computation trees over any infinite subset of a field, the integer RAM, and the generalized real RAM model of Ben-Amram and Galil). Using this technique, we show optimal Ω(n) bounds for dynamic matrix–vector product, dynamic matrix multiplication, and dynamic discriminant and an Ω(n) lower bound for dynamic polynomial multiplication (convolution), providing a good match with Reif and Tate's O(n log n) upper bound. We also show linear lower bounds for dynamic determinant, matrix adjoint, and matrix inverse and an Ω(n) lower bound for the elementary symmetric functions. The second technique is the communication complexity technique of Miltersen, Nisan, Safra, and Wigderson which we apply to the setting of dynamic algebraic problems, obtaining similar lower bounds in the word RAM model. The third technique gives lower bounds in the weaker straight line program model. Using this technique, we show an Ω((log n)2/log log n) lower bound for dynamic discrete Fourier transform. Technical ingredients of our techniques are the incompressibility technique of Ben-Amram and Galil and the lower bound for depth-two superconcentrators of Radhakrishnan and Ta-Shma. The incompressibility technique is extended to arithmetic computation in arbitrary fields.

Discovery of inference rules for question-answering

One of the main challenges in question-answering is the potential mismatch between the expressions in questions and the expressions in texts. While humans appear to use inference rules such as ‘X writes Y’ implies ‘X is the author of Y’ in answering questions, such rules are generally unavailable to question-answering systems due to the inherent difficulty in constructing them. In this paper, we present an unsupervised algorithm for discovering inference rules from text. Our algorithm is based on an extended version of Harris’ Distributional Hypothesis, which states that words that occurred in the same contexts tend to be similar. Instead of using this hypothesis on words, we apply it to paths in the dependency trees of a parsed corpus. Essentially, if two paths tend to link the same set of words, we hypothesize that their meanings are similar. We use examples to show that our system discovers many inference rules easily missed by humans.

Aspects of biodiversity of Scots pine forests in Europe

Management of Scots pine, the most widely distributed conifer in the world, was often based on clearfelling and replanting regimes, resulting in a rather poor biodiversity value. However, there is nowadays a general expectation to increase biodiversity by applying a more complex silviculture. Although present knowledge of genetic factors is insufficient to draw firm conclusions on the desirable level of genetic diversity, it seems unlikely that current silvicultural practices will limit genetic diversity in Scots pine Native pinewoods are rare in Europe, but have a flora and fauna of high conservation value. Therefore they must be recognized as a priority habitat under the European Commission's Habitat Directive. The high conservation value of native stands is a function of their old-growth structure that provides a rare habitat. A number of measures should be taken in all types of Scots pine forests to enhance biodiversity. Firstly, old growth habitats should be promoted. Foresters have to accept that a small percentage of the pine resource should be managed upon much longer rotations. Secondly, retention of deadwood should be encouraged. Dead and dying wood are key components of stand structure and act as key substrates for many associated species, such as microbes, invertebrates, small mammals and birds. Furthermore, a complex stand structure should be promoted at both the horizontal and the vertical scale. A small scale forestry, group regeneration systems, natural regeneration, introduction of broadleaves and stronger thinnings are strongly recommended. Availability of quite precise niches significantly increases biodiversity value. Conservation of isolated populations, found under extreme environmental conditions, is an absolute need. Populations endangered either by their small size or by environmental stresses, hybridization with other species or human interference should be primarily conserved. Forest edges support a range of taxa, and open habitats can comprise many different plant community types. A large number of organisms are directly or indirectly dependent on or favoured by fire. However, enhancing biodiversity provokes also some risks. Generally, browsing is considered as a moderate risk. Introduced species, such as aspen, act as an alternate host to the rust. Open species can present a threat to the European pine marten, dying and deadwood can provide breeding habitats for pest species (Tomicus piniperda), burning increases the risk of seedlings being attacked pine fire fungus (Rhizina undulata) and forest edges may be an attractive habitat for pest insects. An extension of existing growth models is needed to incorporate biodiversity issues in forest management planning. Distance dependent individual tree growth models should be developed. Sets of indicators for biodiversity must integrate compositional, structural and functional attributes. Attributes such as species richness, species abundance, species diversity, horizontal and vertical distribution, tree age, tree size, stand diversity, architectural complexity, genetic variants and deadwood are needed for the establishment of biodiversity indices. Assessment of functional phenomena needs the knowledge of the driving biotic environmental factors.