Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18625 papers
We demonstrate that an unlexicalized PCFG with refined conjunction categories can parse much more accurately than previously shown, by making use of simple, linguistically motivated state splits, which break down false independence assumptions latent in a vanilla treebank grammar and reflect the Chinese idiosyncratic grammatical property. Indeed, its performance is the best result in the 3nd Chinese Parsing Evaluation of single model. This result has showed that refine the function words to represent Chinese subcat frame is a good method. An unlexicalized PCFG is much more compact, easier to replicate, and easier to interpret than more complex lexical models, and the parsing algorithms are simpler, more widely understood, of lower asymptotic complexity, and easier to optimize.
We investigated how shape features in natural images influence emotions aroused in human beings. Shapes and their characteristics such as roundness, angularity, simplicity, and complexity have been postulated to affect the emotional responses of human beings in the field of visual arts and psychology. However, no prior research has modeled the dimensionality of emotions aroused by roundness and angularity. Our contributions include an in-depth statistical analysis to understand the relationship between shapes and emotions. Through experimental results on the International Affective Picture System (IAPS) dataset we provide evidence for the significance of roundness-angularity and simplicitycomplexity on predicting emotional content in images. We combine our shape features with other state-of-theart features to show a gain in prediction and classification accuracy. We model emotions from a dimensional perspective in order to predict valence and arousal ratings which have advantages over modeling the traditional discrete emotional categories. Finally, we distinguish images with strong emotional content from emotionally neutral images with high accuracy.
We investigate the problem of automatically converting from a dependency representation to a phrase structure representation, a key aspect of understanding the relationship between these two representations for NLP work. We implement a new approach to this problem, based on a small number of supertags, along with an encoding of some of the underlying principles of the Penn Treebank guidelines. The resulting system significantly outperforms previous work in such automatic conversion. We also achieve comparable results to a system using a phrase-structure parser for the conversion. A comparison with our system using either the part-of-speech tags or the supertags provides some indication of what the parser is contributing. 1
Learning a foreign language (FL) entails more than attaining the mastery of a system of linguistic norms or the functional and pragmatic aspects of that language. It requires learning to adapt to different cultural norms. So, the challenge is to provide FL learners with opportunities to interact with people from other cultural and linguistic realities and re-focus the aims of FL learning to the development of intercultural communicative competence (ICC). The introduction of information and communication technology (ICT) in education not only enhances the access to information but also enables intercultural contact among individuals from diverse cultural backgrounds, setting the conditions for the development of curriculum-based telecollaboration projects. The LOA eTwinning project presented in this chapter was implemented in the context of an action-research project aimed to introduce an intercultural approach to teaching English to raise pupils’ motivation and challenge them to become more creative, more collaborative, and more autonomous.
This contribution explores the subgroup of text structuring expressions with the form preposition + demonstrative pronoun, thus it is devoted to an aspect of the interaction of coreference relations and relations signaled by discourse connectives (DCs) in a text. The demonstrative pronoun typically signals a referential link to an antecedent, whereas the whole expression can, but does not have to, carry a discourse meaning in sense of discourse connectives. We describe the properties of these phrases/expressions with regard to their antecedents, their position among the text-structuring language means and their features typical for the “connective function ” of them compared to their “non-connective function”. The analysis is carried out on Czech data from the approx. 50,000 sentences of the Prague Dependency Treebank 2.0, directly on the syntactic trees. We explore the characteristics of these phrases/expressions discovered during two projects: the manual annotation of coreference relations (Nedoluzhko et al. 2011) and discourse connectives, their scopes and meanings (Mladová et al. 2008).
This article presents the phonetic and inflection linguistic changes in the three editions of Krzysztof Kluk’s The trees, garden herbs and gardens (edit. 1777, 1797, 1808). The first part of the analysis focuses on those changes that are consistent with the evaluation of the Polish nationwide standardized language at the turn of the eighteenth and nineteenth centuries. The second part of the article presents linguistic changes where the modern and correct forms are changed into the older one, consistent with the north-eastern Polish language of the Eastern Borderlands. Ascertaining why the editors of Kluk’s book used the older linguistic forms instead of the correct forms, leads to attempts to determine the standard linguistic norms in the Piarists’ printing house during the very short period at the turn of the eighteenth and the nineteenth centuries.
The objective of textology (text linguistics) is to analyse and describe interpersonal communication in all its aspects, as it happens now and did in the past, in various discourse communities. The basic categories of so defined textology are: text, utterance and discourse, regarded as different perspectives on the phenomenon of interpersonal communication. This phenomenon manifests itself in established works, with their specific semantic and syntactic structure, but also in genre affinity, in interactive events fixed in a widely understood situational context, and in social, cultural and linguistic norms which regulate communication activities in individual human communities and make interpersonal communication possible. These three categories can be said to reify in definite communication phenomena: in written and recorded text, and in spoken utterances, i.e. in actualized discourses.
"This paper addresses the problem of optimizing the training treebank data because the size and quality of the data has always been a bottleneck for the purposes of training. In previous studies we realized that current corpora used for training machine learning{based dependency parsers contain a significant proportion of redundant information at the syntactic structure level. Since the development of such training corpora involves a big effort, we argue that an appropriate process for selecting the sentences to be included in them can result in having parsing models as accurate as the ones given when training with bigger { non optimized corpora (or alternatively, bigger accuracy for an equivalent annotation effort). This argument is supported by the results of the study we carried out, which is presented in this paper. Therefore, this paper demonstrates that the training corpora contain more information than needed for training accurate data{driven dependency parsers."
This paper presents an identification framework for extracting Tibetan base noun phrase (NP). The framework includes two phases. In the first phase, Chinese base NPs are extracted from all Chinese sentences in the sentence aligned Chinese-Tibetan corpus using Stanford Chinese parser. In the second phase, the Tibetan translations of those Chinese NPs are identified using four different methods, that is, word alignment, iterative re-evaluation, dictionary and word alignment, and sequence intersection method. We implemented and tested these methods on Chinese-Tibetan sentence aligned unlabelled corpus without Tibetan POS tagger and Treebank. The experimental results demonstrate these methods can get satisfactory results, and the best performance with 0.5283 precision is got using sequence intersection identification method. The identification framework can also be extended to extract Tibetan verb phrase.
Informal Mathematical Discourse (IMD) is characterized by the mixture of natural language and symbolic expressions in the context of textbooks, publications in mathematics and mathematical proof. We focused the IMD processing at the low level of discourse. In this paper, we proposed the preprocessing phase before the IMD structure analysis within the context of Controlled Natural Language (CNL). Our contribution is defined in context of the IMD processing and the use of machine learning; first, we present a CNL, a pure corpus and Matemathical Treebank for processing IMD; second, we present a preprocessing phase for IMD analysis with connectives disambiguation and verbs treatment, finally, we found a satisfactory result on input text parsing using a statistical parsing model. We will propagate these results for classification of argumentative informal practices via the low level discourse in IMD processing.
The generalizability and validity of idiographic anger ratings were investigated among female dementia family caregivers (N = 65). The longitudinal (2 weeks) and multi-method measurement approach included standardized measures of anger/hostility, affect ratings elicited by caregiving events, and criticism in audiotaped speech samples. In an application of generalizability theory, variance components of idiographic anger ratings were examined. Ratings of anger elicited by specific caregiving events showed generalizability across time, event, and mood adjectives and were related to verbal criticisms displayed towards the patient and use of psychoactive medications. Intercorrelations among standardized anger measures were quite low. Results from this study indicate a strong need for further measurement development related to the experiences and expressions of anger in family caregivers, and caution by clinicians in selecting tools for practice settings.
Sentence realization, as one of the important components in natural language generation, has taken a statistical swing in recent years. While most previous approaches make heavy usage of lexical information in terms of N-gram language models, we propose a novel method based on unlexicalized tree linearization grammars. We formally define the grammar representation and demonstrate learning from either treebanks with gold-standard annotations, or automatically parsed corpora. For the testing phase, we present a linear time deterministic algorithm to obtain the 1-best word order and further extend it to perform exact search for n-best linearizations. We carry out experiments on various languages and report state-of-the-art performance. In addition, we discuss the advantages of our method on both empirical aspects and its linguistic interpretability.
BACKGROUND: This study examines inequalities in health in Laos. Because perception of health might affect ratings, we used both a global and a relative self-rated health (SRH) question. METHODS: The study was based on the fourth Lao Expenditure and Consumption Survey, 2007-2008. The study population consisted of 24 162 individuals 20 years or older. Two single-question measures of SRH were used: a global with no reference point and a relative with age group reference. RESULTS: Significant associations were found with age, sex, illiteracy, ethnicity, remote location, health measures, nutrition, and household poverty. Worse health was reported using SRH questions with reference points by the young rather than the old. CONCLUSION: In Laos, poor SRH is associated with illiteracy, inaccessibility, Mon-Khmer ethnicity, age, being a woman, and being poor. More factors were found to be associated with global rather than relative SRH.
It is well known that accuracies of statistical parsers trained over Penn treebank on test sets drawn from the same corpus tend to be overestimates of their actual parsing performance. This gives rise to the need for evaluation of parsing performance on corpora from different domains. Evaluating multiple parsers on test sets from different domains can give a detailed picture about the relative strengths/weaknesses of different parsing approaches. Such information is also necessary to guide choice of parser in applications such as machine translation where text from multiple domains needs to be handled. In this paper, we report a benchmarking study of different state-of-art parsers for English, both constituency and dependency. The constituency parser output is converted into CoNLL-style dependency trees so that parsing performance can be compared across formalisms. Specifically, we train rerankers for Berkeley and Stanford parsers to study the usefulness of reranking for handling texts from different domains. The results of our experiments lead to interesting insights about the out-of-domain performance of different English parsers.
Previous joint models of Chinese part-of-speech (POS) tagging and dependency parsing are extended from either graph- or transition-based dependency models. Our analysis shows that the two models have different error distributions. In addition, integration of graph- and transition-based dependency parsers by stacked learning (stacking) has achieved significant improvements. These motivate us to study the problem of stacking graph- and transition-based joint models. We conduct experiments on Chinese Penn Treebank 5.1 (CTB5.1). The results demonstrate that the guided transition-based joint model obtains better performance than the guided graph-based joint model. Further, we introduce a constituent-based joint model which derives the POS tag sequence and dependency tree from the output of PCFG parsers, and then integrate it into the guided transition-based joint model. Finally, we achieve the best performance on CTB5.1, 94.95 % in tagging accuracy and 83.98 % in parsing accuracy respectively.
This paper presents the results of a set of preliminary experiments combining two knowledge-based partial dependency analyzers with two statistical parsers, applied to the Basque Dependency Treebank. The general idea will be to apply a stacked scheme where the output of the rule-based partial parsers will be given as input to MaltParser and MST, two state of the art statistical parsers. The results show a modest improvement over the baseline, although they also present interesting lines for further research.
Due to the explosive growth of the Web, the domain of Web personalization has gained great momentum both in the research and commercial areas. One of the most popular web personalization systems is recommender systems. In recommender systems choosing user information that can be used to profile users is very crucial for user profiling. In Web 2.0, one facility that can help users organize Web resources of their interest is user tagging systems. Exploring user tagging behavior provides a promising way for understanding users’ information needs since tags are given directly by users. However, free and relatively uncontrolled vocabulary makes the user self-defined tags lack of standardization and semantic ambiguity. Also, the relationships among tags need to be explored since there are rich relationships among tags which could provide valuable information for us to better understand users. In this paper, we propose a novel approach for learning tag ontology based on the widely used lexical database WordNet for capturing the semantics and the structural relationships of tags. We present personalization strategies to disambiguate the semantics of tags by combining the opinion of WordNet lexicographers and users’ tagging behavior together. To personalize further, clustering of users is performed to generate a more accurate ontology for a particular group of users. In order to evaluate the usefulness of the tag ontology, we use the tag ontology in a pilot tag recommendation experiment for improving the recommendation performance by exploiting the semantic information in the tag ontology. The initial result shows that the personalized information has improved the accuracy of the tag recommendation.
We consider a specific class of tree structures that can represent basic structures in linguistics and computer science such as XML documents, parse trees, and treebanks, namely, finite node-labeled sibling-ordered trees. We present axiomatizations of the monadic second-order logic (MSO), monadic transitive closure logic (FO(TC1)) and monadic least fixed-point logic (FO(LFP1)) theories of this class of structures. These logics can express important properties such as reachability. Using model-theoretic techniques, we show by a uniform argument that these axiomatizations are complete, i.e., each formula that is valid on all finite trees is provable using our axioms. As a backdrop to our positive results, on arbitrary structures, the logics that we study are known to be non-recursively axiomatizable.
When a video of someone speaking is paused, the stationary image of the speaker typically appears less flattering than the video, which contained motion. We call this the frozen face effect (FFE). Here we report six experiments intended to quantify this effect and determine its cause. In Experiment 1, video clips of people speaking in naturalistic settings as well as all of the static frames that composed each video were presented, and subjects rated how flattering each stimulus was. The videos were rated to be significantly more flattering than the static images, confirming the FFE. In Experiment 2, videos and static images were inverted, and the videos were again rated as more flattering than the static images. In Experiment 3, a discrimination task measured recognition of the static images that composed each video. Recognition did not correlate with flattery ratings, suggesting that the FFE is not due to better memory for particularly distinct images. In Experiment 4, flattery ratings for groups of static images were compared with those for videos and static images. Ratings for the video stimuli were higher than those for either the group or individual static stimuli, suggesting that the amount of information available is not what produces the FFE. In Experiment 5, videos were presented under four conditions: forward motion, inverted forward motion, reversed motion, and scrambled frame sequence. Flattery ratings for the scrambled videos were significantly lower than those for the other three conditions. In Experiment 6, as in Experiment 2, inverted videos and static images were compared with upright ones, and the response measure was changed to perceived attractiveness. Videos were rated as more attractive than the static images for both upright and inverted stimuli. Overall, the results suggest that the FFE requires continuous, natural motion of faces, is not sensitive to inversion, and is not due to a memory effect.
Lexicography in the Arab world has had important effects on the development of the Arabic language. The origin and subsequent development and refinement of traditional Arabic grammatical theory—as early as the eighth century- had intimate links with the practice of writing dictionaries. The Kitab al-ʿayn, by al-Khalil ibn Ahmad (d.c.786), which is the first full-scale dictionary in the Arab world, marked a significant milestone in the history of grammatical thought and set the tone for more works on Arabic grammar. For many centuries, the general mode of the theory has acknowledged a "closed" corpus of Qur'anic diction and pre-Islamic poetry and prose as the major source of Arabic lexicographic works. The main credo is that Arabic dictionaries should contain the "unattained" forms of the language and remain impervious to external persuasions; namely colloquialisms, borrowings, neologisms, and coinages. Arabic dictionaries continued to resist the slightest reform as to the codification of lexical innovations and the treatment of lexical gaps exhausting themselves to the point of stagnation. Today, English-Arabic dictionary editors have to deal with a huge number of lexical gaps that have cumulated over time. The lexical gap-filling process is carried out in a very unsystematic way that is far from creating an atmosphere of cooperation that ultimately contributes to creating unified English-Arabic lexical databases for lexicographic purposes. The paper explains how a modern English-Arabic dictionary can fall short in its modernizing role, and gives a snapshot of the most salient microstructural issues that characterize the Al-Mawrid Al-Hadeeth: A Modern English-Arabic Dictionary (2010).
Traditional Chinese grammar is represented by Li Jinxi's A New Chinese Grammar firstly.Li's grammar system,taking the sentence layout and the sentence component as its main characteristic,is called the sentence-based grammar.This paper firstly briefly reviews the development history of the Chinese grammar,and summarizes the main ideas and theoretical features of two schools: traditional grammar and structural grammar.Then the paper analyzes the advantages and disadvantages of the main grammar systems in Chinese Information Process(CIP) from the view of the Chinese treebank,and compared them to traditional grammar to reveal the necessity of applying traditional grammar to CIP field.Finally,the paper discusses some key issues to be handled in the future of application.
Syntactically annotated corpora have become important resources for natural language processing due in part to the success of corpus-based methods. Since words are often considered as primitive units of language structures, the annotation of word segmentation forms the basis of these corpora. This is also an issue for the Vietnamese Treebank (VTB), which is the first and only publicly available syntactically annotated corpus for the Vietnamese language. Although word segmentation is straight-forward for space-delimited languages like English, this is not the case for languages like Vietnamese for which a standard criterion for word segmentation does not exist. This work explores the challenges of Vietnamese word segmentation through the detection and correction of inconsistency for VTB. Then, by combining and splitting the inconsistent annotations that were detected, we are able to observe the influence of different word segmentation criteria on automatic word segmentation, and the applications of word segmentation, including text classification and English-Vietnamese statistical machine translation. The analysis and experimental results showed that our methods improved the quality of VTB, which positively affected the performance of its applications. Title and Abstract in another language, L2 (optional, and on same page) So sánh các tiêu chí tách từ khác nhau thông qua ứng dụng
Annotation of discourse relations is a project related to the Prague Dependency Treebank 2.5. It represents a new manually annotated layer of language description, above the existing layers of the PDT, and it portrays linguistic phenomena from the perspective of discourse structure and coherence.
MT systems typically use parsers to help reorder constituents. However most languages do not have adequate treebank data to learn good parsers, and such training data is extremely time-consuming to annotate. Our earlier work has shown that a reordering model learnt from word-alignments using POS tags as features can improve MT performance (Visweswariah et al., 2011). In this paper, we investigate the effect of word-classing on reordering performance using this model. We show that unsupervised word clusters perform somewhat worse but still reasonably well, compared to a part-of-speech (POS) tagger built with a small amount of annotated data; while a richer tagset including case and gender-number-person further improves reordering performance by around 1.2 monolingual BLEU points. While annotating this richer tagset is more complicated than annotating the base tagset, it is much easier than annotating treebank data. Keywords:SMT, Reordering, POS-Tagging 1.
A neurocomputational model based on emergent massively overlapping neural cell assemblies (CAs) for resolving prepositional phrase (PP) attachment ambiguity is described. PP attachment ambiguity is a well-studied task in natural language processing and is a case where semantics is used to determine the syntactic structure. A large network of biologically plausible fatiguing leaky integrate-and-fire neurons is trained with semantic hierarchies (obtained from WordNet) on sentences with PP attachment ambiguity extracted from the Penn Treebank corpus. During training, overlapping CAs representing semantic similarities between the component words of the ambiguous sentences emerge and then act as categorizers for novel input. The resulting average resolution accuracy of 84.56% is on par with known machine learning algorithms.
We present a Bayesian nonparametric model for estimating tree insertion grammars (TIG), building upon recent work in Bayesian inference of tree substitution grammars (TSG) via Dirichlet processes. Under our general variant of TIG, grammars are estimated via the Metropolis-Hastings algorithm that uses a context free grammar transformation as a proposal, which allows for cubic-time string parsing as well as tree-wide joint sampling of derivations in the spirit of Cohn and Blunsom (2010). We use the Penn treebank for our experiments and find that our proposal Bayesian TIG model not only has competitive parsing performance but also finds compact yet linguistically rich TIG representations of the data. 1
In this investigation the paper uses Condi-tional Random Field(CRF) for automatic identification of Concurrent Structure in Chinese Infor-mation Processing.We explores the distribution of Concurrent Structure in Tsinghua University 973 Treebank and analyze the internal and external linguistic features of the structure.A condi-tional random field model is built using these features.In open test,the best F value reaches 85.71% and this result shows that CRF is an effective method for automatic identification of Concurrent structure.
Previous experiments suggest that a rule-based approach to tree alignment error correction serves to be an eective complement to statistical alignment. We show how, using relatively few features, an implementation of Brill’s Transformation-Based Learning algorithm improves the results of a high precision model of the statistical aligner Lingua-Align. Using our system to correct already tree aligned data, we achieve balanced F-scores of 80.6 on our test set and 85.2 on our development test set. Using it as a tree aligner on word aligned data, our best F-scores using the same model amount to 78.7 and 83.0 respectively. Finally, we apply a pipeline of alignment and error correction tools to create several versions of a large parallel treebank consisting of various domains for Dutch to English for use in a syntax-based MT system. We conclude that transformation-based learning is a promising approach for the large-scale creation of parallel treebanks for various NLP purposes.
We provide a model that extends the split-merge framework of Petrov et al. (2006) to jointly learn latent annotations and Tree Sub-stitution Grammars (TSGs). We then conduct a variety of experiments with this model, first inducing grammars on a portion of the Penn Treebank and the Korean Treebank 2.0, and next experimenting with grammar refinement from a single nonterminal and from the Uni-versal Part of Speech tagset. We present quali-tative analysis showing promising signs across all experiments that our combined approach successfully provides for greater flexibility in grammar induction within the structured guidance provided by the treebank, leveraging the complementary natures of these two ap-proaches. 1
This article discusses how linguistic and translation norms, as evident in dictionaries, enforce the ideology of heteronormativity in Slovenia. The aim of this paper is to show how translation norms concerning homoeroticism were shaped in the translation of classical literature in Slovenia in the twentieth century. Translation norms are shaped in a certain period of time and in a certain environment among translators and others involved in translation according to social and cultural circumstances, expectations, and adaptations of topics to these expectations, in which the translation contrasts the initial norms with the norms of the target culture. At the same time, the linguistic norms of the standard language are created as a result of speakers’ continuous adaptation to a social and cultural environment, as a result of adapting to the current social ideal. It is assumed that translations contributed to creating a model of heteronormativity, which continues to characterize the Slovenian community today because of the limited number of new translations of classical works of literature. The paper concludes with a brief analysis of evidence of homosexuality in Slovenian translations of Shakespeare’s The Merchant of Venice.
Overview Since the early nineties, the on-going dramatic loss of the world’s linguistic diversity has gained attention, first by the linguists and increasingly also by the general public. As a response, the new field of language documentation emerged from around 2000 on, starting with the funding initiative ‘Dokumentation Bedrohter Sprachen’ (DoBeS, funded by the Volkswagen foundation, Germany), soon to be followed by others such as the ‘Endangered Languages Documentation Programme’ (ELDP, at SOAS, London), or, in the USA, ‘Electronic Meta-structure for Endangered Languages Documentation’ (EMELD, led by the LinguistList) and ‘Documenting Endangered Languages’ (DEL, by the NSF). From its very beginning, the new field focused on digital technologies not only for recording in audio and video, but also for annotation, lexical databases, corpus building and archiving, among others. This development not just coincides but is intrinsically interconnected with the increasing focus on digital data, technology and methods in all sciences, in particular in the humanities.
A novel method for hybrid graph-based dependency parsing of natural language text is proposed. It is based on k-best maximum spanning tree dependency parsing and evaluation of the spanning trees by using a verb valency lexicon for a given language as a reranking knowledge base. The approach is compared with existing state-of-the-art transition-based and graph-based approaches to dependency parsing. As the proposed generic method was developed specifically for improving the accuracy of Croatian dependency parsing, Croatian Dependency Treebank and CROVALLEX verb valency lexicon are used in the experiment. The suggested approach scored approximately 77.21% LAS, outperforming the tested state-of-the-art approaches by at least 2.68% LAS.
A morphological analyser only recognizes words that it already knows in the lexical database. It needs, however, a way of sensing significant changes in the language in the form of newly borrowed or coined words with high frequency. We develop a finite-state morphological guesser in a pipelined methodology for extracting unknown words, lemmatizing them, and giving them a priority weight for inclusion in a lexicon. The processing is performed on a large contemporary corpus of 1,089,111,204 words and passed through a machine-learning-based annotation tool. Our method is tested on a manually-annotated gold standard of 1,310 forms and yields good results despite the complexity of the task. Our work shows the usability of a highly non-deterministic finite state guesser in a practical and complex application. 1
This thesis studies parsing and literature with the Data-Oriented Parsing framework, which assumes that chunks of previous experience can be exploited to analyze new sentences. As chunks we consider syntactic tree fragments. After presenting a method to efficiently extract such fragments from treebanks based on heuristics of re-occurrence, we employ them to develop a multi-lingual statistical parser. We show how a mildly context-sensitive grammar can be employed to produce discontinuous constituents, and compare this to an approximation that stays within the efficiently parsable context-free framework. We show that tree fragments allow the grammar to adequately capture the statistical regularities of non-local relations, without the need for the increased generative capacity of mildly context-sensitive grammar. The second part investigates what separates literary from other novels. We work with a corpus of novels and a reader survey with ratings of how literary they are perceived to be. The main goal is to find out the extent to which the literary ratings can be predicted from the texts. We first evaluate simple measures such as vocabulary richness, text compressibility, and the number of cliché expressions. In addition we apply more sophisticated, predictive models: a topic model, bag-of-words model, and a model based on syntactic tree fragments. We find that literary ratings are predictable from textual features to a large extent. While it is not possible to infer a causal relation, this result clearly rules out the notion that these value-judgments of literary merit were arbitrary, or predominantly determined by factors beyond the text.
This is an exploratory study investigating potential effects of emotional valence in images and their influence on conversation in the presence of the images. We used latent-semantic analysis to generalize valence ratings of Swedish words to a corpus of spoken conversations. Each utterance in the conversation was given a valence rating, which represented how emotionally positive or emotionally negative the utterance was. We found no effects that indicate that valenced images have an effect on conversations. However, we find that valenced images in general, and positive images in particular, were considered more helpful by the participants who engaged in the conversations. Additionally, we find no results that interlocutors align over time in their use of valenced language.
Syntax-based translation models that operate on the output of a source-language parser have been shown to perform better if allowed to choose from a set of possible parses. In this paper, we investigate whether this is because it allows the translation stage to overcome parser errors or to override the syntactic structure itself. We find that it is primarily the latter, but that under the right conditions, the translation stage does correct parser errors, improving parsing accuracy on the Chinese Treebank. 1
This thesis introduces an experimental and quantitative approach to language through the study of the concept of soft constraints and its application to two phenomena of order in French: the position of the attributive adjective and the ordering of verbal complements occurring in postverbal position. Soft constraints are defined as affecting the acceptability rather than the grammaticality of the sentences. Our main hypothesis is that these constraints are properties of the language and thus must studied in syntax. These constraints raise a methodological issue: since they do not affect the grammaticality of the sentences, they cannot be investigated using the traditional tools of syntax (introspection and grammaticality judgment). It is therefore necessary to define tools for their description and analysis. The proposed methods are statistical analysis of corpus date, inspired by the work of Bresnan et al. (2007) and Bresnan & Ford (2010) and, to a lesser extent, psycholinguistic experiment. Regarding the position of the adjective, we test most of the constraints encountered in the literature and we propose a statistical analysis of the data extracted from the French Treebank corpus. We show the importance of the adjectival item and the nominal item with which it combines. Other constraints linked to the internal syntax of the adjectival phrase and the noun phrase also play a significant role in the choice of position. The work on the relative order of the verbal complements is conducted on a sample of sentences extracted from two newspapers corpora (French Treebank and Est-Républicain) and two corpora of spoken French (ESTER and C-ORAL-ROM). We show the significant influence of the constituent weight over the ordering: short before long order which is a feature of SVO languages like French, is observed in over 86% of cases. We also identify the important role of the verbal lemma associated with its semantic class (annotated with the dictionary of Dubois & Dubois-Charlier, 1997). Finally, building on the analysis of corpus data as well a two questionnaires eliciting acceptability judgments, it seems that nor animacy neither information structure (given/new, Prince, 1981) have a significant effect on the postverbal complement ordering.
Impairment in social communication skills is one of the core deficits in the children with Autism Spectrum Disorder (ASD). These children are characterized by an inherent inability to express their affective states thereby imposing limitations on traditional self-report and observational methodologies. However physiological signals are continuously available and are arguably not impacted by these difficulties. In recent years several assistive technologies utilizing the benefits of physiology-based systems have been investigated to promote social communication skills in this population. Among these we chose Virtual Reality (VR) as our platform. Investigations in the area of Human Computer Interaction (HCI) have shown that variations in the physiological signals can be evoked by different amounts of presence in the VR environment and the transition from one affective state to another is accompanied by dynamic shift in indicators of Autonomic Nervous System (ANS) activity. The presented work seeks to fuse behavioral viewing pattern and peripheral physiological features with the affective rating. Thus, this is a step towards indicating the potential of such a system to build an intelligent therapist-like affect-recognizer. The preliminary findings of a usability study are promising.
Syntactic analysis is a valuable addition to corpus annotation, since it enables the retrieval of structural information which is otherwise difficult to access. The Norwegian Infrastructure for the Exploration of Syntax and Semantics is adding syntactic information to the Norwegian Newspaper Corpus and is effectively producing a treebank as a parsed corpus.
The chapter presents a meta-search tool developed in order to deliver search results structured according to the specific interests of users. Meta-search means that for a specific query, several search mechanisms could be simultaneously applied. Using the clustering process, thematically homogenous groups are built up from the initial list provided by the standard search mechanisms. The results are more user oriented, as a result of the ontological approach of the clustering process. After the initial search made on multiple search engines, the results are pre-processed and transformed into vectors of words. These vectors are mapped into vectors of concepts, by calling an educational ontology and using the WordNet lexical database. The vectors of concepts are refined through concept space graphs and projection mechanisms, before applying the clustering procedure. Implementation details and early experimentation results are also provided.
This thesis explores the development of parallel treebanks, collections of language data consisting of texts and their translations, with syntactic annotation and alignment, linking words, phrases, and sentences to show translation equivalence. We describe the semi-manual annotation of the SMULTRON parallel treebank, consisting of 1,000 sentences in English, German and Swedish. This description is the starting point for answering the first of two questions in this thesis. What issues need to be considered to achieve a high-quality, consistent,parallel treebank? The units of annotation and the choice of annotation schemes are crucial for quality, and some automated processing is necessary to increase the size. Automatic quality checks and evaluation are essential, but manual quality control is still needed to achieve high quality. Additionally, we explore improving the automatically created annotation for one language, using information available from the annotation of the other languages. This leads us to the second of the two questions in this thesis. Can we improve automatic annotation by projecting information available in the other languages? Experiments with automatic alignment, which is projected from two language pairs, L1–L2 and L1–L3, onto the third pair, L2–L3, show an improvement in precision, in particular if the projected alignment is intersected with the system alignment. We also construct a test collection for experiments on annotation projection to resolve prepositional phrase attachment ambiguities. While majority vote projection improves the annotation, compared to the basic automatic annotation, using linguistic clues to correct the annotation before majority vote projection is even better, although more laborious. However, some structural errors cannot be corrected by projection at all, as different languages have different wording, and thus different structures.
This paper outlines the design principles and choices, as well as the ongoing development process of the Common Orthographic Vocabulary of the Portuguese Language (VOC), a large scale electronic lexical database which was adopted by the Community of Portuguese-Speaking Countries ’ (CPLP) Instituto Internacional da Língua Portuguesa to implement a spelling reform that is currently taking place. Given the different available resources and lexicographic traditions within the CPLP countries, a range of different solutions was adopted for different countries and integrated into a common development framework. Although the publication of lexicographic resources to implement spelling reforms has always been done for Portuguese, VOC represents a paradigm change, switching from idiosyncratic, closed source, paper-format official resources to standardized, open, free, web-accessible and reusable ones. We start by outlining the context that justifies the resource development and its requirements, then focusing on the description of the methodology, workflow and tools used, showing how a collaborative project in a common web-based platform and administration interface make the creation of such a long-sought and ambitious project possible.
The GF Eclipse Plugin provides an integrated development environment (IDE) for developing grammars in the Grammatical Framework (GF). Built on top of the Eclipse Platform, it aids grammar writing by providing instant syntax checking, semantic warnings and cross-reference resolution. Inline documentation and a library browser facilitate the use of existing resource libraries, and compilation and testing of grammars is greatly improved through single-click launch configurations and an in-built test case manager for running treebank regression tests. This IDE promotes grammar-based systems by making the tasks of writing grammars and using resource libraries more efficient, and provides powerful tools to reduce the barrier to entry to GF and encourage new users of the framework.
We present results of a study investigating evaluative learning in dementia patients with a classic evaluative conditioning paradigm. Picture pairs of three unfamiliar faces with liked, disliked, or neutral faces, that were rated prior to the presentation, were presented 10 times each to a group of dementia patients (N = 15) and healthy controls (N = 14) in random order. Valence ratings of all faces were assessed before and after presentation. In contrast to controls, dementia patients changed their valence ratings of unfamiliar faces according to their pairing with either a liked or disliked face, although they were not able to explicitly assign the picture pairs after the presentation. Our finding suggests preserved evaluative conditioning in dementia patients. However, the result has to be considered preliminary, as it is unclear which factors prevented the predicted rating changes in the expected direction in the control group.
Chinese word structure annotation is potentially useful for many NLP tasks, especially for Chinese word segmentation. Li and Zhou (2012) have presented an annotation for word structures in the Penn Chinese Treebank. But they only consider words that have productive affixes, which covers 35% of word types in that corpus. In this paper, we propose a linguistically inspired annotation that covers various morphological derivations of Chinese in a more general way, such that almost all multiple-character words can be structurally analyzed. As manual annotation is expensive, we propose a semi-supervised approach to automatic annotation, which combines the maximum entropy learning and the EM iteration for the Gaussian mixture model. The proposed method has achieved an accuracy of 90% on the testing set. © 2021 CLP 2012 - 2nd CIPS-SIGHAN Joint Conference on Chinese Language Processing. All Rights Reserved.
A novel statistical linguistic feature, called punctuation confidence, is proposed in this paper for assisting in prosodic break prediction in Mandarin text-to-speech. The punctuation confidence calculated from the input text is a measure of the likelihood of inserting a major PM at a word boundary. Since a punctuation in text tends to be pronounced as a break, the punctuation confidence associated with a punctuation estimate should provide useful information for break prediction from text. The idea is realized in this study by first employing a conditional random field (CRF)-based model to generate a predicted punctuation and its associated punctuation confidence for each word boundary. Then, the predicted punctuation and its punctuation confidence are combined with contextual linguistic features to predict the break type of the word boundary by an MLP (multi-layer perceptrons). Experiment on the Treebank speech corpus confirmed the effectiveness of the proposed approach.
Texts The Prague Czech-English Dependency Treebank 2.0 (PCEDT 2.0) is a major update of the Prague Czech-English Dependency Treebank 1.0 (LDC2004T25). It is a manually parsed Czech-English parallel corpus sized over 1.2 million running words in almost 50,000 sentences for each part. Data The English part contains the entire Penn Treebank - Wall Street Journal Section (LDC99T42). The Czech part consists of Czech translations of all of the Penn Treebank-WSJ texts. The corpus is 1:1 sentence-aligned. An additional automatic alignment on the node level (different for each annotation layer) is part of this release, too. The original Penn Treebank-like file structure (25 sections, each containing up to one hundred files) has been preserved. Only those PTB documents which have both POS and structural annotation (total of 2312 documents) have been translated to Czech and made part of this release. Each language part is enhanced with a comprehensive manual linguistic annotation in the PDT 2.0 style (LDC2006T01, Prague Dependency Treebank 2.0). The main features of this annotation style are: dependency structure of the content words and coordinating and similar structures (function words are attached as their attribute values) semantic labeling of content words and types of coordinating structures argument structure, including an argument structure ("valency") lexicon for both languages ellipsis and anaphora resolution. This annotation style is called tectogrammatical annotation and it constitutes the tectogrammatical layer in the corpus. For more details see below and documentation. Annotation of the Czech part Sentences of the Czech translation were automatically morphologically annotated and parsed into surface-syntax dependency trees in the PDT 2.0 annotation style. This annotation style is sometimes called analytical annotation; it constitutes the analytical layer of the corpus. The manual tectogrammatical (deep-syntax) annotation was built as a separate layer above the automatic analytical (surface-syntax) parse. A sample of 2,000 sentences was manually annotated on the analytical layer. Annotation of the English part The resulting manual tectogrammatical annotation was built above an automatic transformation of the original phrase-structure annotation of the Penn Treebank into surface dependency (analytical) representations, using the following additional linguistic information from other sources: PropBank (LDC2004T14) VerbNet NomBank (LDC2008T23) flat noun phrase structures (by courtesy of D. Vadas and J.R. Curran) For each sentence, the original Penn Treebank phrase structure trees are preserved in this corpus together with their links to the analytical and tectogrammatical annotation.
WordNet proved that it is possible to construct a large-scale electronic lexical database on the principles of lexical semantics. It has been accepted and used extensively by computational linguists ever since it was released. Inspired by WordNet's success, we propose as an alternative a similar resource, based on the 1987 Penguin edition of Roget's Thesaurus of English Words and Phrases. Peter Mark Roget published his first Thesaurus over 150 years ago. Countless writers, orators and students of the English language have used it. Computational linguists have employed Roget's for almost 50 years in Natural Language Processing, however hesitated in accepting Roget's Thesaurus because a proper machine tractable version was not available. This dissertation presents an implementation of a machine-tractable version of the 1987 Penguin edition of Roget's Thesaurus - the first implementation of its kind to use an entire current edition. It explains the steps necessary for taking a machine-readable file and transforming it into a tractable system. This involves converting the lexical material into a format that can be more easily exploited, identifying data structures and designing classes to computerize the Thesaurus. Roget's organization is studied in detail and contrasted with WordNet's. We show two applications of the computerized Thesaurus: computing semantic similarity between words and phrases, and building lexical chains in a text. The experiments are performed using well-known benchmarks and the results are compared to those of other systems that use Roget's, WordNet and statistical techniques. Roget's has turned out to be an excellent resource for measuring semantic similarity; lexical chains are easily built but more difficult to evaluate. We also explain ways in which Roget's Thesaurus and WordNet can be combined.