Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
18265 papers
A critical task in Affective Computing is the reliable assessment of emotional states. The two most prominent approaches to classify emotions are categorical concepts of discrete emotions (e.g. OCC) and dimensional models typically using the pleasure - arousal - dominance space (PAD). In current research and applications, however, there is little overlap between these two concepts. A mapping of discrete categories into the dimensional space would offer new possibilities to model the emotional states of users and artificial agents, though. We hence let N=70 healthy subjects place the labels of discrete OCC emotions into PAD space according to their subjective knowledge with a simple visual tool. There was a high inter-subject consistency regarding the positioning of OCC emotions for the dimension of pleasure. However, arousal and dominance ratings showed considerably greater variance. We conclude that global and reliable mappings of OCC emotions into the PAD space can best be provided for the pleasure dimension. The exact positioning of discrete emotions regarding arousal and dominance can only be gained by individual calibration of a given user in a strict within-subject approach.
There has recently been some interest among computational linguists in the task of inducing grammar-based "semantic parsers" from sets of paired strings and meaning representations, following pioneering work by Zettlemoyer and Collins (2005). Work of this kind is currently limited by the paucity of datasets for training. The talk reviews the state of the art in this field, then proposes a way to semi-automatically generate much larger language-independent datasets, on the same order of magnitude as syntactic treebanks, using linguistic knowledge that has only recently begun to become available, for use in inducing semantic parsers for under-resourced languages for application in statistical machine translation.
Syntactic representations based on word-to-word dependencies have a long-standing tradition in descriptive linguistics, and receive considerable interest in many applications. Nevertheless, dependency syntax has remained something of an island from a formal point of view. Moreover, most formalisms available for dependency grammar are restricted to projective analyses, and thus not able to support natural accounts of phenomena such as wh-movement and cross–serial dependencies. In this article we present a formalism for non-projective dependency grammar in the framework of linear context-free rewriting systems. A characteristic property of our formalism is a close correspondence between the non-projectivity of the dependency trees admitted by a grammar on the one hand, and the parsing complexity of the grammar on the other. We show that parsing with unrestricted grammars is intractable. We therefore study two constraints on non-projectivity, block-degree and well-nestedness. Jointly, these two constraints define a class of “mildly” non-projective dependency grammars that can be parsed in polynomial time. An evaluation on five dependency treebanks shows that these grammars have a good coverage of empirical data.
Cost-sensitive classification, where the features used in machine learning tasks have a cost, has been explored as a means of balancing knowledge against the expense of incrementally obtaining new features. We introduce a setting where humans engage in classification with incrementally revealed features: the collegiate trivia circuit. By providing the community with a web-based system to practice, we collected tens of thousands of implicit word-by-word ratings of how useful features are for eliciting correct answers. Observing humans' classification process, we improve the performance of a state-of-the art classifier. We also use the dataset to evaluate a system to compete in the incremental classification task through a reduction of reinforcement learning to classification. Our system learns when to answer a question, performing better than baselines and most human players.
BACKGROUND: Previous studies addressing teaching and learning in point-of-care ultrasound have primarily focussed on image interpretation and not on the technical quality of the images. We hypothesized that a limited intervention of 10 supervised examinations would improve the technical skills in Focus Assessed Transthoracic Echocardiography (FATE) and that physicians with no experience in FATE would quickly adopt technical skills allowing for image quality suitable for interpretation. METHODS: Twenty-one physicians with no previous training in FATE or echocardiography (Novices) participated in the study and a reference group of three examiners with more than 10 years of experience in echocardiography (Experts) was included. Novices received an initial theoretical and practical introduction (2 hours), after which baseline examinations were performed on two healthy volunteers. Subsequently all physicians were scheduled to a separate intervention day comprising ten supervised FATE examinations. For effect measurement a second examination (evaluation) of the same two healthy volunteers from the baseline examination was performed. RESULTS: At baseline 86% of images obtained by novices were suitable for interpretation, on evaluation this was 93% (p = 0.005). 100% of images obtained by experts were suitable for interpretation. Mean global image rating on baseline examinations was 70.2 (CI 68.0-72.4) and mean global image rating after intervention was 75.0 (CI 72.9-77.0), p = 0.0002. In comparison, mean global image rating in the expert group was 89.8 (CI 88.8-90.9). CONCLUSIONS: Improvement of technical skills in FATE can be achieved with a limited intervention and upon completion of intervention 93% of images achieved are suitable for clinical interpretation.
Este artículo describe investigación sobre los efectos de la desambiguación morfosintáctica usada como un preproceso de un analizador sintáctico profundo basado en\nHPSG, en el contexto del desarrollo de un treebank del español de código abierto, en el\nentorno de DELPH-IN. La anotación treebank se realiza manualmente tomando las decisiones\napropiadas entre las opciones propuestas por el sistema y ordenadas por un módulo\nestadístico. Los experimentos presentados muestran que el uso de un etiquetador reduce\nla ambigüedad de las frases, y contribuye a limitar la cantidad de frases cuyo análisis sobrepasa el límite de tiempo, y ayuda a al módulo estadístico a clasificar el árbol correcto entre los mejores. Por un lado, nuestros resultados validan los beneficios ya reportados en la literatura de tal preproceso de análisis profundo con respecto a la velocidad, cobertura y precisión. Por otro lado, proponemos una estrategia basada en existentes herramientas de código abierto y recursos para desarrollar con alta consitencia treebanks de sintaxis profunda\npara idiomas con limitada disponibilidad de recursos lingüísticos.
Recent study shows that parsing accuracy can be largely improved by the joint optimization of part-of-speech (POS) tagging and dependency parsing. However, the POS tagging task does not benefit much from the joint framework. We argue that the fundamental reason behind is because the POS features are overwhelmed by the syntactic features during the joint optimization, and the joint models only prefer such POS tags that are favourable solely from the parsing viewpoint. To solve this issue, we propose a separately passive-aggressive learning algorithm (SPA), which is designed to separately update the POS features weights and the syntactic feature weights under the joint optimization framework. The proposed SPA is able to take advantage of previous joint optimization strategies to significantly improve the parsing accuracy, but also overcome their shortages to significantly boost the tagging accuracy by effectively solving the syntax-insensitive POS ambiguity issues. Experiments on the Chinese Penn Treebank 5.1 (CTB5) and the English Penn Treebank (PTB) demonstrate the effectiveness of our proposed methodology and empirically verify our observations as discussed above. We achieve the best tagging and parsing accuracies on both datasets, 94.60% in tagging accuracy and 81.67 % in parsing accuracy on CTB5, and 97.62 % and 93.52 % on PTB.
We aimed to investigate whether fear of suffocation predicts healthy persons' respiratory and affective responses to obstructed breathing as evoked by inspiratory resistive loads. Participants (N = 27 women, ages between 18 and 21 years) completed the Fear of Suffocation scale and underwent 16 trials in which an inspiratory resistive load of 15 cmH(2)O/l/s (small) or 40 cmH(2)O/l/s (large) was added to the breathing circuit for 40 s. Fear of suffocation was associated with higher arousal ratings for both loads. Loaded breathing was associated with a decrease in minute ventilation, but progressively less so for participants scoring higher on fear of suffocation when breathing against the large load. The present findings document a potentially panicogenic mechanism that may maintain and worsen respiratory discomfort in persons with fear of suffocation.
There is growing evidence that drugs of abuse alter processing of emotional information in ways that could be attractive to users. Our recent report that Δ⁹-tetrahydrocannabinol (THC) diminishes amygdalar activation in response to threat-related faces suggests that THC may modify evaluation of emotionally-salient, particularly negative or threatening, stimuli. In this study, we examined the effects of acute THC on evaluation of emotional images. Healthy volunteers received two doses of THC (7.5 and 15 mg; p.o.) and placebo across separate sessions before performing tasks assessing facial emotion recognition and emotional responses to pictures of emotional scenes. THC significantly impaired recognition of facial fear and anger, but it only marginally impaired recognition of sadness and happiness. The drug did not consistently affect ratings of emotional scenes. THC's effects on emotional evaluation were not clearly related to its mood-altering effects. These results support our previous work, and show that THC reduces perception of facial threat. Nevertheless, THC does not appear to positively bias evaluation of emotional stimuli in general.
Effects of a Protein Optimized Diet Combined with Moderate Resistance Training on the Postoperative Course in Older Patients with Hip Fracture
Employing higher-order subtree structures in graph-based dependency parsing has shown substantial improvement over the accuracy, however suffers from the inefficiency increasing with the order of subtrees. We present a new reranking approach for dependency parsing that can utilize complex subtree representation by applying efficient subtree selection heuristics. We demonstrate the effective-ness of the approach in experiments conducted on the Penn Treebank and the Chinese Treebank. Our system improves the baseline accuracy from 91.88 % to 93.37 % for English, and in the case of Chinese from 87.39 % to 89.16%. 1.
From the perspective of structural linguistics, we explore paradigmatic and syntagmatic lexical relations for Chinese POS tagging, an important and challenging task for Chinese language processing. Paradigmatic lexical relations are explicitly captured by word clustering on large-scale unlabeled data and are used to design new features to enhance a discriminative tagger. Syntagmatic lexical relations are implicitly captured by constituent parsing and are utilized via system combination. Experiments on the Penn Chinese Treebank demonstrate the importance of both paradigmatic and syntagmatic relations. Our linguistically motivated approaches yield a relative error reduction of 18 % in total over a stateof-the-art baseline. 1
Linear Context-Free Rewriting System (LCFRS) is an extension of Context-Free Grammar (CFG) in which a non-terminal can dominate more than a single continu-ous span of terminals. Probabilistic LCFRS have recently successfully been used for the direct data-driven parsing of discontin-uous structures. In this paper we present a parser for binary PLCFRS of fan-out two, together with a novel monotonous estimate for A ∗ parsing, with which we conduct ex-periments on modified versions of the Ger-man NeGra treebank and the Discontinuous Penn Treebank in which all trees have block degree two. The experiments show that compared to previous work, our approach provides an enormous speed-up while de-livering an output of comparable richness. 1
According to an influential dual-process model, a moral judgment is the outcome of a rapid, affect-laden process and a slower, deliberative process. If these outputs conflict, decision time is increased in order to resolve the conflict. Violations of deontological principles proscribing the use of personal force to inflict intentional harm are presumed to elicit negative affect which biases judgments early in the decision-making process. This model was tested in three experiments. Moral dilemmas were classified using (a) decision time and consensus as measures of system conflict and (b) the aforementioned deontological criteria. In Experiment 1, decision time was either unlimited or reduced. The dilemmas asked whether it was appropriate to take a morally questionable action to produce a "greater good" outcome. Limiting decision time reduced the proportion of utilitarian ("yes") decisions, but contrary to the model's predictions, (a) vignettes that involved more deontological violations logged faster decision times, and (b) violation of deontological principles was not predictive of decisional conflict profiles. Experiment 2 ruled out the possibility that time pressure simply makes people more like to say "no." Participants made a first decision under time constraints and a second decision under no time constraints. One group was asked whether it was appropriate to take the morally questionable action while a second group was asked whether it was appropriate to refuse to take the action. The results replicated that of Experiment 1 regardless of whether "yes" or "no" constituted a utilitarian decision. In Experiment 3, participants rated the pleasantness of positive visual stimuli prior to making a decision. Contrary to the model's predictions, the number of deontological decisions increased in the positive affect rating group compared to a group that engaged in a cognitive task or a control group that engaged in neither task. These results are consistent with the view that early moral judgments are influenced by affect. But they are inconsistent with the view that (a) violation of deontological principles are predictive of differences in early, affect-based judgment or that (b) engaging in tasks that are inconsistent with the negative emotional responses elicited by such violations diminishes their impact.
OBJECTIVE: To develop a Spanish version of the WHO-Composite International Diagnostic Interview (WHO-CIDI) applicable to Spain, through cultural adaptation of its most recent Latin American (LA v 20.0) version. METHODS: A 1-week training course on the WHO-CIDI was provided by certified trainers. An expert panel reviewed the LA version, identified words or expressions that needed to be adapted to the cultural or linguistic norms for Spain, and proposed alternative expressions that were agreed on through consensus. The entire process was supervised and approved by a member of the WHO-CIDI Editorial Committee. The changes were incorporated into a Computer Assisted Personal Interview (CAPI) format and the feasibility and administration time were pilot tested in a convenience sample of 32 volunteers. RESULTS: A total of 372 questions were slightly modified (almost 7% of approximately 5000 questions in the survey) and incorporated into the CAPI version of the WHO-CIDI. Most of the changes were minor - but important - linguistic adaptations, and others were related to specific Spanish institutions and currency. In the pilot study, the instrument's mean completion administration time was 2h and 10min, with an interquartile range from 1.5 to nearly 3h. All the changes made were tested and officially approved. CONCLUSIONS: The Latin American version of the WHO-CIDI was successfully adapted and pilot-tested in its computerized format and is now ready for use in Spain.
The majority of fear conditioning studies in humans have focused on fear acquisition rather than fear extinction. For this reason only a few functional imaging studies on fear extinction are available. A large number of animal studies indicate the medial prefrontal cortex (mPFC) as neuronal substrate of extinction. We therefore determined mPFC contribution during extinction learning after a discriminative fear conditioning in 34 healthy human subjects by using functional near-infrared spectroscopy. During the extinction training, a previously conditioned neutral face (conditioned stimulus, CS+) no longer predicted an aversive scream (unconditioned stimulus, UCS). Considering differential valence and arousal ratings as well as skin conductance responses during the acquisition phase, we found a CS+ related increase in oxygenated haemoglobin concentration changes within the mPFC over the time course of extinction. Late CS+ trials further revealed higher activation than CS- trials in a cluster of probe set channels covering the mPFC. These results are in line with previous findings on extinction and further emphasize the mPFC as significant for associative learning processes. During extinction, the diminished fear association between a former CS+ and a UCS is inversely correlated with mPFC activity--a process presumably dysfunctional in anxiety disorders.
and determiner errors with spelling correction as a pre-processing step. The result shows that spelling correction improves the Detection, Correction, and Recognition F-scores for preposition errors. With regard to preposition error correction, F-scores were not improved when using the training set with correction of all but preposition errors. As for determiner error correction, there was an improvement when the constituent parser was trained with a concatenation of treebank and modified treebank where all the articles appearing as the first word of an NP were removed. Our system ranked third in preposition and fourth in determiner error corrections. 1
Drawing from social identity theory, this research examines scarce gender representation as a contextual condition that inhibits same‐gender supervisors' support. Survey results in S tudy 1 found that when women were proportionally underrepresented, they reported feeling less supported by female supervisors than male supervisors. S tudy 2 showed that women who perceived they were gender tokens in their organization were less likely to support an outstanding female subordinate than an identical male. S tudy 3 experimentally tested social mobility as a mechanism for the effects of tokenism on same‐gender supervisor support. Results suggest that social mobility and group composition jointly affect ratings of same‐gender targets. Perceptions of gender‐based social mobility appear to be one mechanism through which tokenism influences same‐gender relations at work.
This paper presents a theoretical discussion about the use of Frame Semantics as corpora annotation paradigm. The objective of this paper is to evaluate the applicability of Frame Semantics theory and FrameNet paradigm for the semantic annotation of legal texts. The work presented in this paper is an initial step in the construction of a treebank for the Brazilian legal language.
The common use of a single de facto standard annotation scheme for dependency treebank creation leaves the question open to what extent the performance of an application trained on a treebank depends on this annotation scheme and whether a linguistically richer scheme would imply a decrease of the performance of the application. We investigate the effect of the variation of the number of grammatical relations in a tagset on the performance of dependency parsers. In order to obtain several levels of granularity of the annotation, we design a hierarchical annotation scheme exclusively based on syntactic criteria. The richest annotation contains 60 relations. The more coarse-grained annotations are derived from the richest. As a result, all annotations and thus also the performance of a parser trained on different annotations remain comparable. We carried out experiments with four state-of-the-art dependency parsers. The results support the claim that annotating with more fine-grained syntactic relations does not necessarily imply a significant loss of accuracy. We also show the limits of this approach by giving details on the fine-grained relations that do have a negative impact on the performance of the parsers.
The paper concentrates on which language means may be included into the annotation of discourse relations in the Prague Dependency Treebank (PDT) and tries to examine the so called alternative lexicalizations of discourse markers (AltLex’s) in Czech. The analysis proceeds from the annotated data of PDT and tries to draw a comparison between the Czech AltLex’s from PDT and English AltLex’s from PDTB (the Penn Discourse Treebank). The paper presents a lexico-syntactic and semantic characterization of the Czech AltLex’s and comments on the current stage of their annotation in PDT. In the current version, PDT contains 306 expressions (within the total 43,955 of sentences) that were labeled by annotators as being an AltLex. However, as the analysis demonstrates, this number is not final. We suppose that it will increase after the further elaboration, as AltLex’s are not restricted to a limited set of syntactic classes and some of them exhibit a great degree of variation. Key words: alternative lexicalization of discourse markers (AltLex); discourse connectives; discourse relations 1.
the effect of pathological aging on explicit memory is very well documented, but relatively few studies have addressed this issue in the musical domain. To examine learning and consolidation of melodies, we designed a melodic recognition task involving immediate and delayed recognition of 16 target melodies (8 familiar and 8 unfamiliar). Seventeen patients with mild to moderate Alzheimer's disease (AD) and 17 age-matched controls were tested. During the initial presentation of the targets, the participant had to decide whether or not the melody was familiar. Recognition was tested after one and three presentations of the target melodies using a yes/no recognition paradigm. Delayed recognition was tested after 24 hours to evaluate consolidation. In keeping with the findings of Bartlett, Halpern, and Dowling (1995), age-matched controls showed better recognition of familiar than unfamiliar melodies. Controls also showed improved performance with multiple presentations for both familiar and unfamiliar melodies, without forgetting after 24-hour delay. In contrast, patients with AD showed impaired learning and recognition of both unfamiliar and familiar melodies with no benefit of familiarity on recognition. Nevertheless, the familiarity decision-based ratings of patients was in keeping with controls. These findings suggest that musical recognition memory is impaired in AD, but the musical lexicon (as assessed by familiarity ratings) is preserved. These findings highlight the need to use both familiar and unfamiliar music in experimental tasks to study the different processes underlying recognition memory.
Statistical machine translation has been remarkably successful for the world’s well-resourced languages, and much effort is focussed on creating and exploiting rich resources such as treebanks and wordnets. Machine translation can also support the urgent task of documenting the world’s endangered languages. The primary object of statistical translation models, bilingual aligned text, closely coincides with interlinear text, the primary artefact collected in documentary linguistics. It ought to be possible to exploit this similarity in order to improve the quantity and quality of documentation for a language. Yet there are many technical and logistical problems to be addressed, starting with the problem that – for most of the languages in question – no texts or lexicons exist. In this position paper, we examine these challenges, and report on a data collection effort involving 15 endangered languages spoken in the highlands of
The paper presents a small empirical study into emotion and affect recognition based on auditory and visual features, which was performed in the context of the Audio-Visual Emotion Challenge (AVEC) 2012. The goal of this competition is to predict continuous-valued affect ratings based on the provided auditory and visual features, e.g., local binary pattern (LBP) features extracted from aligned face images, and spectral audio features.
The paper analyzes 33 grammatical structures of Tsinghua University 973 Treebank from the syntactic point of view. Firstly, we explore the distribution of these structures in Tsinghua University 973 Treebank and analyze the syntactic constituents of these structures. Then, we gather statistics about these structures based on the external functional relation and the internal structural relation of its subcomponents. Finally, according to the same internal structural relations, we generate a matrix to show the ambiguity among these structures. The statistical data offers the syntactic knowledge for auto-identifying these structures in future use.
BACKGROUND: Previous studies on nitrogen narcosis have focused on how it affects behavior, performance, and cognitive function. However, little is known about the effects of nitrogen narcosis on the emotional processing of external stimuli. METHOD: We presented 20 volunteers with images from the International Affective Picture System (IAPS) and categorized as unpleasant, neutral, or pleasant, while sitting in a hyperbaric chamber at the surface (101,3kPa) and at 39 m equivalent depth (496.4 kPa). The participants rated the images along three affective dimensions: valence (intrinsic attractiveness or aversiveness of a stimuli), arousal, and dominance. RESULTS: In the valence dimension there was no significant effect of increased pressure or interaction between increased pressure and image category. There was a significant interaction between image category and the pressure at which the images were viewed in the arousal dimension. The mean arousal rating score for unpleasant stimuli was 0.5 point (on a 9-point scale) lower at hyperbaric conditions and equal arousal rating score for neutral stimuli in general. DISCUSSION: The absence of any effect of pressure in the valence dimension suggests that divers have no impairment in their ability to determine the pleasantness or unpleasantness of different stimuli. Furthermore, this study suggests that the effects of nitrogen narcosis on the emotional processing of external stimuli are primarily evident in the arousal dimension. Although differences in arousal ratings were relatively small in magnitude, even a small alteration in emotional response to external stimuli might be important in the context of deep diving.
While our knowledge about ancient civilizations comes mostly from studies in archaeology and history books, much can also be learned or confirmed from literary texts. Using natural language processing techniques, we present aspects of ancient China as revealed by statistical textual analysis on the Complete Tang Poems, a 2.6-million-character corpus of all surviving poems from the Tang Dynasty (AD 618—907). Using an automatically created treebank of this corpus, we outline the semantic profiles of various poets, and discuss the role of s easons, geography, history, architecture, and colours, as observed through word selection and dependencies.
Impaired risk recognition has been suggested to be associated with the risk for revictimization and the development of posttraumatic stress disorder (PTSD). Moreover, risk behavior has been linked to high sensation seeking, which may also increase the probability of revictimization. A newly designed behavioral experiment with five audiotaped risk scenarios was used to investigate risk recognition in revictimized, single-victimized, and nontraumatized individuals with and without PTSD. Moreover, the potential role of sensation seeking in revictimization, and PTSD as well as its relation to risk recognition was explored. Revictimized, single-victimized, and nontraumatized individuals did not differ with regard to general risk recognition. However, delayed risk recognition was found for the revictimized group when arousal ratings were considered. No differences in sensation seeking were found between the three groups; only the nontraumatized group showed lower boredom susceptibility relative to the revictimized group. Delayed risk recognition was associated with high sensation seeking. Furthermore, PTSD symptoms significantly predicted exit levels of risk scenarios. Findings are discussed against the background of previous research.
OBJECTIVE: To prospectively compare an indocyanine green (ICG)-enhanced optical imaging system with contrast-enhanced magnetic resonance imaging (MRI) for the detection of synovitis in the hands of patients with rheumatologic disorders. METHODS: Forty-five patients (30 women [67%], mean ± SD age 52.6 ± 13.4 years) in whom there was a clinical suspicion of an inflammatory arthropathy were examined with a commercially available device for ICG-enhanced optical imaging as well as by contrast-enhanced 3T MRI as the standard of reference. Three independent readers graded the degree of synovitis in the carpal, metacarpophalangeal, proximal interphalangeal, and distal interphalangeal joints of both hands (1,350 joints), using a 4-point ordinate scale (0 = no synovitis, 1 = mild, 2 = moderate, 3 = severe). Statistical analyses were performed using a logistic generalized estimating equation approach. Agreement of optical imaging ratings made by the different readers was estimated with a weighted kappa coefficient. RESULTS: When MRI was used as the standard of reference, optical imaging showed a sensitivity of 39.6% (95% confidence interval [95% CI] 31.1-48.7%), a specificity of 85.2% (95% CI 79.5-89.5%), and accuracy of 67.0% (95% CI 61.4-72.1%) for the detection of synovitis in patients with arthritis. Diagnostic accuracy was especially limited in the setting of mild synovitis, while it was substantially better in patients with severely inflamed joints. Moderate interreader and intrareader agreement was observed. CONCLUSION: The evaluated ICG-enhanced optical imaging system showed limitations for the detection of inflamed joints of the hand in comparison with MRI.
OBJECTIVE: Relation extraction in biomedical text mining systems has largely focused on identifying clause-level relations, but increasing sophistication demands the recognition of relations at discourse level. A first step in identifying discourse relations involves the detection of discourse connectives: words or phrases used in text to express discourse relations. In this study supervised machine-learning approaches were developed and evaluated for automatically identifying discourse connectives in biomedical text. MATERIALS AND METHODS: Two supervised machine-learning models (support vector machines and conditional random fields) were explored for identifying discourse connectives in biomedical literature. In-domain supervised machine-learning classifiers were trained on the Biomedical Discourse Relation Bank, an annotated corpus of discourse relations over 24 full-text biomedical articles (~112,000 word tokens), a subset of the GENIA corpus. Novel domain adaptation techniques were also explored to leverage the larger open-domain Penn Discourse Treebank (~1 million word tokens). The models were evaluated using the standard evaluation metrics of precision, recall and F1 scores. RESULTS AND CONCLUSION: Supervised machine-learning approaches can automatically identify discourse connectives in biomedical text, and the novel domain adaptation techniques yielded the best performance: 0.761 F1 score. A demonstration version of the fully implemented classifier BioConn is available at: http://bioconn.askhermes.org.
Affect is increasingly recognized as central to decision making. However, it is not clear whether affect can be used to predict choice. To address this issue, we conducted 4 studies designed to create and test a model that could predict choice from affect. In Study 1, we used an image rating task to develop a model that predicted approach-avoidance motivations. This model quantified the role of two basic dimensions of affect--valence and arousal--in determining choice. We then tested the predictive power of this model for two types of decisions involving images: preference based selections (Study 2) and risk-reward trade-offs (Study 3). In both cases, the model derived in Study 1 predicted choice and outperformed competing models drawn from well-established theoretical views. Finally, we showed that this model has ecological validity: It predicted choices between news articles on the basis of headlines (Study 4). These findings have implications for diverse fields, including neuroeconomics and judgment and decision making.
Slang is one of the problems encountered in developing Lexical Database. There is no complete source officially for slangs in any language. This research attempts to create a list of Indonesian slangs using the proposed framework and methodology. The contents for slangs are generated by Twitter; therefore it can encounter the newest slangs available in the society.
In the following paper, we discuss and evaluate the benefits that deep syntactic trees (tectogrammatics) and all the rich annotation of the Prague Dependency Treebank bring to the process of annotating the discourse structure, i.e. discourse relations, connectives and their arguments. The decision to annotate discourse structure directly on the trees contrasts with the majority of similarly aimed projects, usually based on the annotation of linear texts. Our basic assumption is that some syntactic features of a sentence analysis correspond to certain discourselevel features. Hence, we use some properties of the dependency-based large-scale treebank of Czech to help establish an independent annotation layer of discourse. The question that we answer in the paper is how much did we gain by employing this approach. TITLE AND ABSTRACT IN CZECH Pomaha tektogramatika při anotaci diskurznich vztahů?
PCFGs can grow exponentially as additional annotations are added to an initially simple base grammar. We present an approach where multiple annotations coexist, but in a factored manner that avoids this combinatorial explosion. Our method works with linguisticallymotivated annotations, induced latent structure, lexicalization, or any mix of the three. We use a structured expectation propagation algorithm that makes use of the factored structure in two ways. First, by partitioning the factors, it speeds up parsing exponentially over the unfactored approach. Second, it minimizes the redundancy of the factors during training, improving accuracy over an independent approach. Using purely latent variable annotations, we can efficiently train and parse with up to 8 latent bits per symbol, achieving F1 scores up to 88.4 on the Penn Treebank while using two orders of magnitudes fewer parameters compared to the naïve approach. Combining latent, lexicalized, and unlexicalized annotations, our best parser gets 89.4 F1 on all sentences from section 23 of the Penn Treebank. 1
In the present paper, we describe in detail and evaluate the process of semi-automatic annotation of intra-sentential discourse relations in the Prague Dependency Treebank, which is a part of the project of otherwise mostly manual annotation of all (intra- and inter-sentential) discourse relations with explicit connectives in the treebank. Our assumption that some syntactic features of a sentence analysis (in a form of a deepsyntax dependency tree) correspond to certain discourse-level features proved to be correct, and the rich annotation of the treebank allowed us to automatically detect the intra-sentential discourse relations, their connectives and arguments in most of the cases. TITLE AND ABSTRACT IN CZECH Poloautomatická anotace vnitrovětných diskurzních vztahů v PDT ABSTRAKT V tomto článku nabízíme detailní popis a evaluaci procesu poloautomatické anotace vnitrovětných textových vztahů v Pražském závislostním korpusu jako součást projektu jinak především manuální anotace všech (vnitro- a mezivětných) textových vztahů s explicitním konektorem v tomto korpusu. Potvrdil se náš předpoklad, že některé syntaktické vlastnosti analýzy věty (ve formě závislostního stromu hloubkové syntaxe) odpovídají jistým vlastnostem na úrovni analýzy textových vztahů (diskurzu). Bohatá anotace korpusu nám ve většině případů umožnila automaticky detekovat vnitrovětné vztahy, jejich konektory a argumenty.
Implicit discourse relation classification is a challenge task due to missing discourse connective. Some work directly adopted machine learning algorithms and linguistically informed features to address this task. However, one interesting solution is to automatically predict implicit discourse connective. In this paper, we present a novel two-step machine learning-based approach to implicit discourse relation classification. We first use machine learning method to automatically predict the discourse connective that can best express the implicit discourse relation. Then the predicted implicit discourse connective is used to classify the implicit discourse relation. Experiments on Penn Discourse Treebank 2.0 (PDTB) and Biomedical Discourse Relation Bank (BioDRB) show that our method performs better than the baseline system and previous work.
OBJECTIVES: Understanding the relationship between the menstrual cycle and pain can contribute significantly to our knowledge of pain processing in women. Many early studies suggested that pain sensitivity was enhanced during the luteal phase of the menstrual cycle relative to the follicular phase; however, these studies were often limited by small sample sizes, lack of ovulation verification, focus on a single pain modality, inadequate assessment of menstrual cycle regularity, and low-powered statistical methods. The current study was designed to address these limitations and examine the difference in pain processing between the mid-follicular (days 5 to 8) and late-luteal (days 1 to 6 preceding menses) phases. METHODS: Forty-one healthy, regularly cycling women attended testing sessions that measured pain sensitivity from mechanical pain threshold, electrocutaneous pain threshold/tolerance, and ischemia pain threshold/tolerance, as well as McGill Pain Questionnaire qsensory and affective ratings of electric and ischemic stimuli. Electrocutaneous stimulation was also used to assess nociceptive flexion reflex threshold, a physiological measure of spinal nociception. RESULTS: When analyses were limited to data collected only in the targeted menstrual phases (N=30), results indicated no menstrual phase effect on any pain outcome (all P's>0.05), with the exception of lower electrocutaneous pain thresholds during the late-luteal phase. No outcomes differed by menstrual phase in the full sample (N=41). This indicates nociceptive responding varies little between the mid-follicular and late-luteal phases. DISCUSSION: The present study suggests that experimental pain processing does not significantly differ between the mid-follicular and late-luteal phases of the menstrual cycle in healthy women. This implies hormonal variation across these 2 phases (ie, progesterone) has a minimal effect on subjective and physiological responses to pain.
International audience
The basic colour terms for black and white are studied in four archaic and two contemporary linguistic norms of the Chinese language. It is presented that studied Chinese linguistic norms use a common term for white and three different terms for black. It is suggested that the different basic colour terms for black might originate from different source languages. The study supports a panchronic language development instead of a diachronic one, and includes introductions to histories of the Chinese linguistic norms
We demonstrate a novel, robust vision-tolanguage generation system called Midge. Midge is a prototype system that connects computer vision to syntactic structures with semantic constraints, allowing for the automatic generation of detailed image descriptions. We explain how to connect vision detections to trees in Penn Treebank syntax, which provides the scaffolding necessary to further refine data-driven statistical generation approaches for a variety of end goals. 1
OBJECTIVE: It has traditionally been thought that covert face recognition cannot be observed in developmental cases of prosopagnosia, because the phenomenon is thought to rely on the activation of face representations created during a period of normal processing. Yet, recent studies have provided evidence of covert recognition in some developmental cases, and critically the findings of one study suggest that these individuals might be processing faces on an affective dimension rather than a familiarity dimension. The current study aimed to examine this possibility using a physiological measure of covert recognition, the skin conductance response (SCR). METHOD: One 61-year-old male with developmental prosopagnosia and 10 age-matched (M = 59.80 years, SD = 4.02) controls (5 men) took part in this study. Participants viewed a set of 15 famous faces intermixed with 30 novel faces, and the SCR was recorded throughout. RESULTS: Although control participants demonstrated an increased SCR for famous faces in comparison with novel faces, t(9) = 2.112, p =.032, d =.382, the same finding was not observed in Patient WS. However, when WS' increase in SCR was correlated with his affective ratings of the celebrities from name cues, a strong negative correlation was observed (r = -.614, n = 34, p =.020). CONCLUSION: This pattern of findings was interpreted as evidence that WS is covertly processing faces on an affective dimension rather than a familiarity dimension, and fits well with recent neurophysiological findings that support hypotheses for independent processing of cognitive and affective information.
The Quran is a significant religious text written in a unique literary style, close to very poetic language in nature. Accordingly it is significantly richer and more complex than the newswire style used in the previously released Arabic PropBank (Zaghouani et al., 2010; Diab et al., 2008). We present preliminary work on the creation of a unique Arabic proposition repository for Quranic Arabic. We annotate the semantic roles for the 50 most frequent verbs in the Quranic Arabic Dependency Treebank (QATB) (Dukes and Buckwalter 2010). The Quranic Arabic PropBank (QAPB) will be a unique new resource of its kind for the Arabic NLP research community as it will allow for interesting insights into the semantic use of classical Arabic, poetic literary Arabic, as well as significant religious texts. Moreover, on a pragmatic level QAPB will add approximately 810 new verbs to the existing Arabic PropBank (APB). In this pilot experiment, we leverage our knowledge and experience from our involvement in the APB project. All the QAPB annotations will be made freely available for research purposes. 1
This study investigated the effect of arousal on short-term relational memory and its underlying cortical network. Seventeen healthy participants performed a picture by location, short-term relational memory task using emotional pictures. Functional magnetic resonance imaging was used to measure the blood-oxygenation-level dependent signal relative to task. Subjects' own ratings of the pictures were used to obtain subjective arousal ratings. Subjective arousal was found to have a dose-dependent effect on activations in the prefrontal cortex, amygdala, hippocampus, and in higher order visual areas. Serial position analyses showed that high arousal trials produced a stronger primacy and recency effect than low arousal trials. The results indicate that short-term relational memory may be facilitated by arousal and that this may be modulated by a dose-response function in arousal-driven neuronal regions.