Papers reviewed and determined not to be word norm studies. Use the flag icon to report errors or suggest re-inclusion.
16504 papers
The Web has accumulated a rich source of information, such as text, image, rating, etc, which represent different aspects of user preferences. However, the heterogeneous nature of this information makes it difficult for recommender systems to leverage in a unified framework to boost the performance. Recently, the rapid development of representation learning techniques provides an approach to this problem. By translating the various information sources into a unified representation space, it becomes possible to integrate heterogeneous information for informed recommendation.
Comunicacio presentada a: the Fourth International Conference on Dependency Linguistics (Depling 2017), celebrada a Pisa, Italia, del 18 al 20 de setembre 2017.
Virtual reality (VR) has been proposed as a methodological tool to study the basic science of psychology and other fields. One key advantage of VR is that sharing of virtual content can lead to more robust replication and representative sampling. A database of standardized content will help fulfill this vision. There are two objectives to this study. First, we seek to establish and allow public access to a database of immersive VR video clips that can act as a potential resource for studies on emotion induction using virtual reality. Second, given the large sample size of participants needed to get reliable valence and arousal ratings for our video, we were able to explore the possible links between the head movements of the observer and the emotions he or she feels while viewing immersive VR. To accomplish our goals, we sourced for and tested 73 immersive VR clips which participants rated on valence and arousal dimensions using self-assessment manikins. We also tracked participants' rotational head movements as they watched the clips, allowing us to correlate head movements and affect. Based on past research, we predicted relationships between the standard deviation of head yaw and valence and arousal ratings. Results showed that the stimuli varied reasonably well along the dimensions of valence and arousal, with a slight underrepresentation of clips that are of negative valence and highly arousing. The standard deviation of yaw positively correlated with valence, while a significant positive relationship was found between head pitch and arousal. The immersive VR clips tested are available online as supplemental material.
We describe a variant of Child-Sum Tree-LSTM deep neural network (Tai et al, 2015) fine-tuned for working with dependency trees and morphologically rich languages using the example of Polish. Fine-tuning included applying a custom regularization technique (zoneout, described by (Krueger et al., 2016), and further adapted for Tree-LSTMs) as well as using pre-trained word embeddings enhanced with sub-word information (Bojanowski et al., 2016). The system was implemented in PyTorch and evaluated on phrase-level sentiment labeling task as part of the PolEval competition.
We systematically explore regularizing neural networks by penalizing low\nentropy output distributions. We show that penalizing low entropy output\ndistributions, which has been shown to improve exploration in reinforcement\nlearning, acts as a strong regularizer in supervised learning. Furthermore, we\nconnect a maximum entropy based confidence penalty to label smoothing through\nthe direction of the KL divergence. We exhaustively evaluate the proposed\nconfidence penalty and label smoothing on 6 common benchmarks: image\nclassification (MNIST and Cifar-10), language modeling (Penn Treebank), machine\ntranslation (WMT'14 English-to-German), and speech recognition (TIMIT and WSJ).\nWe find that both label smoothing and the confidence penalty improve\nstate-of-the-art models across benchmarks without modifying existing\nhyperparameters, suggesting the wide applicability of these regularizers.\n
In this paper, I will describe the methodology to develop the first sample of a dependency treebank collecting German aesthetic writings of the late 18th century. A gold standard of the target data was annotated in order to evaluate some datadriven tools, trained on contemporary web news. Results are reported and discussed.
Large-scale distributed training requires significant communication bandwidth for gradient exchange that limits the scalability of multi-node training, and requires expensive high-bandwidth network infrastructure. The situation gets even worse with distributed training on mobile devices (federated learning), which suffers from higher latency, lower throughput, and intermittent poor connections. In this paper, we find 99.9% of the gradient exchange in distributed SGD is redundant, and propose Deep Gradient Compression (DGC) to greatly reduce the communication bandwidth. To preserve accuracy during compression, DGC employs four methods: momentum correction, local gradient clipping, momentum factor masking, and warm-up training. We have applied Deep Gradient Compression to image classification, speech recognition, and language modeling with multiple datasets including Cifar10, ImageNet, Penn Treebank, and Librispeech Corpus. On these scenarios, Deep Gradient Compression achieves a gradient compression ratio from 270x to 600x without losing accuracy, cutting the gradient size of ResNet-50 from 97MB to 0.35MB, and for DeepSpeech from 488MB to 0.74MB. Deep gradient compression enables large-scale distributed training on inexpensive commodity 1Gbps Ethernet and facilitates distributed training on mobile. Code is available at: https://github.com/synxlin/deep-gradient-compression.
With the advent of word embeddings, lexicons are no longer fully utilized for sentiment analysis although they still provide important features in the traditional setting. This paper introduces a novel approach to sentiment analysis that integrates lexicon embeddings and an attention mechanism into Convolutional Neural Networks. Our approach performs separate convolutions for word and lexicon embeddings and provides a global view of the document using attention. Our models are experimented on both the SemEval'16 Task 4 dataset and the Stanford Sentiment Treebank and show comparative or better results against the existing state-of-the-art systems. Our analysis shows that lexicon embeddings allow building high-performing models with much smaller word embeddings, and the attention mechanism effectively dims out noisy words for sentiment analysis.
Discourse parsing has long been treated as a stand-alone problem independent from constituency or dependency parsing. Most attempts at this problem are pipelined rather than end-to-end, sophisticated, and not self-contained: they assume gold-standard text segmentations (Elementary Discourse Units), and use external parsers for syntactic features. In this paper we propose the first end-to-end discourse parser that jointly parses in both syntax and discourse levels, as well as the first syntacto-discourse treebank by integrating the Penn Treebank with the RST Treebank. Built upon our recent span-based constituency parser, this joint syntacto-discourse parser requires no preprocessing whatsoever (such as segmentation or feature extraction), achieves the state-of-the-art end-to-end discourse parsing accuracy.
Recurrent neural networks (RNNs), such as long short-term memory networks (LSTMs), serve as a fundamental building block for many sequence learning tasks, including machine translation, language modeling, and question answering. In this paper, we consider the specific problem of word-level language modeling and investigate strategies for regularizing and optimizing LSTM-based models. We propose the weight-dropped LSTM which uses DropConnect on hidden-to-hidden weights as a form of recurrent regularization. Further, we introduce NT-ASGD, a variant of the averaged stochastic gradient method, wherein the averaging trigger is determined using a non-monotonic condition as opposed to being tuned by the user. Using these and other regularization strategies, we achieve state-of-the-art word level perplexities on two data sets: 57.3 on Penn Treebank and 65.8 on WikiText-2. In exploring the effectiveness of a neural cache in conjunction with our proposed model, we achieve an even lower state-of-the-art perplexity of 52.8 on Penn Treebank and 52.0 on WikiText-2.
Because of the importance of the information conveyed by the clinical documents and owing to the large quantity of raw texts produced in the healthcare system, it became a determinant challenge, in the NLP research field, to arrange the extraction and the management of meaningful data, starting from real text occurrences. In this paper we approach a corpus of 5000 medical diagnoses with sophisticated linguistic and computational devices, which are able to access the semantic dimension of words and sentences contained in it. Our morphosemantic method is grounded on a list of neoclassical formative elements pertaining to the medical domain which has been used for the automatic creation and population of medical lexical resources. The outcomes of this work are automatically built electronic dictionaries and thesauri and an annotated corpus for the NLP in the medical domain.
Because of the importance of the information conveyed by the clinical documents and owing to the large quantity of raw texts produced in the healthcare system, it became a determinant challenge, in the NLP research field, to arrange the extraction and the management of meaningful data, starting from real text occurrences. In this paper we approach a corpus of 5000 medical diagnoses with sophisticated linguistic and computational devices, which are able to access the semantic dimension of words and sentences contained in it. Our morphosemantic method is grounded on a list of neoclassical formative elements pertaining to the medical domain which has been used for the automatic creation and population of medical lexical resources. The outcomes of this work are automatically built electronic dictionaries and thesauri and an annotated corpus for the NLP in the medical domain.
This paper presents Autobank, a prototype tool for constructing a widecoverage Minimalist Grammar (MG) The front end of the tool is a graphical user interface which facilitates the rapid development of a seed set of MG trees via manual reannotation of PTB preterminals with MG lexical categories. The system then extracts various dependency mappings between the source and target trees, and uses these in concert with a non-statistical MG parser to automatically reannotate the rest of the corpus. Autobank thus enables deep treebank conversions (and subsequent modifications) without the need for complex transduction algorithms accompanied by cascades of ad hoc rules; instead, the locus of human effort falls directly on the task of grammar construction itself.
For the data-driven approach in Natural Language Processing (NLP) applications, good quality linguistic resources considered as a main factor to obtain good results. Although Arabic language is one of the main languages in the world, it is considered as low-resourced language in term of good quality and free linguistic resources. This work presents the first stage of building a new open source dependency treebank for Arabic language. It describes the prototype of the new dependency treebank that are inspired by Lexical Functional Grammar (LFG). This paper shows a main approach of developing a newly treebank and put lines the future work needed to complete this novel linguistic resource.
The Universal Dependencies Project1\n(Nivre, [9]; Nivre et al., [10]) is an\nongoing effort towards creating a set of harmonised dependency treebanks\nthat are annotated and structured according to universal guidelines. This paper reports on the addition of morphological features to the Irish Universal\nDependencies Treebank (IUDT). Our feature set subscribes to the feature inventory of the UD Project and has been mapped from Irish morpho-syntactic\ntags – the output of a Finite State Morphological Analyser for Irish (Uí Dhonnchadha and van Genabith [16]). Irish, a Celtic language, has some relatively unusual morphological features that require language-specific labels\nnot covered by the universal feature set. In this paper, we summarise the\nIrish-specific features that we have added to this set by explaining the linguistic properties that they each describe. We also report on the first parsing\nexperiments using the IUDT by assessing the effect that the inclusion of morphological features has on parsing accuracy.
The amount of data that is available for research grows rapidly, yet technology to efficiently interpret and excavate these data lags behind. For instance, when using large treebanks for linguistic research, the speed of a query leaves much to be desired. GrETEL Indexing, or GrInding, tackles this issue. The idea behind GrInding is to make the search space as small as possible before actually starting the treebank search, by pre-processing the treebank at hand. We recursively divide the treebank into smaller parts, called subtree-banks, which are then converted into database files. All subtree-banks are organized according to their linguistic dependency pattern, and labeled as such. Additionally, general patterns are linked to more specific ones. By doing so, we create millions of databases, and given a linguistic structure we know in which databases that structure can occur, leading up to a significant efficiency boost. We present the results of a benchmark experiment, testing the effect of the GrInding procedure on the SoNaR-500 treebank.
Multiword expressions (MWEs) are linguistic objects containing two or more words and showing idiosyncratic behavior at different levels. Treebanks with annotated MWEs enable studies of such properties, as well as training and evaluation of MWE-aware parsers. However, few treebanks contain full-fledged MWE annotations. We show how this gap can be bridged in Polish by projecting 3 MWE resources on a constituency treebank.
Abstract This paper addresses the feasibility of cross-lingual parsing with Universal Dependencies (UD) between Romance languages, analyzing its performance when compared to the use of manually annotated resources of the target languages. Several experiments take into account factors such as the lexical distance between the source and target varieties, the impact of delexicalization, the combination of different source treebanks or the adaptation of resources to the target language, among others. The results of these evaluations show that the direct application of a parser from one Romance language to another reaches similar labeled attachment score (LAS) values to those obtained with a manual annotation of about 3,000 tokens in the target language, and unlabeled attachment score (UAS) results equivalent to the use of around 7,000 tokens, depending on the case. These numbers can noticeably increase by performing a focused selection of the source treebanks. Furthermore, the removal of the words in the training corpus (delexicalization) is not useful in most cases of cross-lingual parsing of Romance languages. The lessons learned with the performed experiments were used to build a new UD treebank for Galician, with 1,000 sentences manually corrected after an automatic cross-lingual annotation. Several evaluations in this new resource show that a cross-lingual parser built with the best combination and adaptation of the source treebanks performs better (77 percent LAS and 82 percent UAS) than using more than 16,000 (for LAS results) and more than 20,000 (UAS) manually labeled tokens of Galician.
In this paper, we describe a method for mapping the phonological feature location of Swedish Sign Language (SSL) signs to the meanings in the Swedish semantic dictionary SALDO.By doing so, we observe clear differences in the distribution of meanings associated with different locations on the body.The prominence of certain locations for specific meanings clearly point to iconic mappings between form and meaning in the lexicon of SSL, which pinpoints modalityspecific properties of the visual modality.
The paper contains the research of noun-compounds from modern Tibetan corpus with the use of a relational lexical database. The lexical database represents a consistent classification of meanings of Tibetan lexical units with different relations between them. The paper describes the structure of the database; principles of process work with Tibetan compounds; recognized types of compounds semantic structure.
In this paper, we propose using a "bootstrapping" method for constructing a dependency treebank of Arabic tweets. This method uses a rule-based parser to create a small treebank of one thousand Arabic tweets and a data-driven parser to create a larger treebank by using the small treebank as a seed training set. We are able to create a dependency treebank from unlabelled tweets without any manual intervention. Experiments results show that this method can improve the speed of training the parser and the accuracy of the resulting parsers.
In this paper, we present evaluation of URDU.KON-TB in the dependency parsing domain. The URDU.KON-TB treebank is developed on the bases of the phrase structure and hyper dependency structure which are only functional constituent's label. Treebank was annotated with three levels of annotation tagset, the semi-semantic POS (SSP), semi-semantic Syntactic (SSS) and Functional (F) tagset and was checked for the Phrase Structure Parsing domain. To evaluate this treebank in the Dependency Parsing domain we have selected MaltParser. To use data in the parser, we have converted the URDU.KON-TB treebank annotated data according to the CONLL format. The compatibility of data to CoNLL is also measured along with usability of data in the dependency parsing domain. To make the data compatible, few assumptions are taken. The converted data is used to evaluate the system by dividing 80% data as training data and 20% data as testing data. We have performed eight experiments. Four experiments are conducted with six different feature models with converted data. The experiments results show URDU.KON-TB treebank is not suitable for the dependency parsing as dependency relation because Head information was missing in the treebank. We then performed four experiments with an assumption based enhancement by adding Head information. The algorithm used to train and test data is Nivre arc-agear algorithm. The new experiments show this treebank data can be used to develop new dependency treebank for Urdu.
This paper describes the use of GrETEL for linguistic research. GrETEL is a linguistic search tool that enables users to look up constructions in syntactically annotated corpora or <i>treebanks</i>. It provides online access to the data, allowing users to query a treebank using either an example sentence or an XPath expression in order to look for similar constructions. A major asset of GrETEL is that it enables non-technical users to consult treebanks in a user-friendly way, which is also in line with the main CLARIN goal of applying the results of speech and language technology to research in the humanities and the social sciences. Besides a description of the querying procedure in GrETEL, this paper presents a selection of research in Dutch syntax and semantics that has been carried out using GrETEL. Furthermore, an overview is given of further developments.
The paper illustrates an effective and innovative method for detecting erroneously annotated arcs in gold dependency treebanks based on an algorithm originally developed to measure the reliability of automatically produced dependency relations. The method permits to significantly restrict the error search space and, more importantly, to reliably identify patterns of systematic recurrent errors which represent dangerous evidence to a parser which tendentially will replicate them. Achieved results demonstrate effectiveness and reliability of the method.
Public textual cyberbullying has become one of the most prevalent issues associated with\nonline safety of young people, particularly on social networks. To address this issue, we\nargue that the boundaries of what constitutes public textual cyberbullying needs to be first\nidentified and a corresponding linguistically motivated definition needs to be advanced.\nThus, we propose a definition of public textual cyberbullying that contains three necessary\nand sufficient elements: the personal marker, the dysphemistic element and the\ncyberbullying link between the previous two elements. Subsequently, we argue that one of\nthe cornerstones in the overall process of mitigating the effects of cyberbullying is the\ndesign of a cyberbullying lexical database that specifies what linguistic and cyberbullying\nspecific information is relevant to the detection process. In this vein, we propose a novel\ncyberbullying lexical database based on the definition of public textual cyberbullying. The\noverall architecture of our cyberbullying lexical database is determined semantically, and, in\norder to facilitate cyberbullying detection, the lexical entry encapsulates two new semantic\ndimensions that are derived from our definition: cyberbullying function and cyberbullying\nreferential domain. In addition, the lexical entry encapsulates other semantic and syntactic \ninformation, such as sense and syntactic category, information that, not only aids the\nprocess of detection, but also allows us to expand the cyberbullying database using\nWordNet (Miller, 1993).
Similarity calculation between business process models has an important role in managing repository of business process model. One of its uses is to facilitate the searching process of models in the repository. Business process similarity is closely related to semantic string similarity. Semantic string similarity is usually performed by utilizing a lexical database such as WordNet to find the semantic meaning of the word. The activity name of the business process uses terms that specifically related to the business field. However, most of the terms in business domain are not available in WordNet. This case would decrease the semantic analysis quality of business process model. Therefore, this study would try to improve semantic analysis of business process model. We present a new lexical database called B-BabelNet. B-BabelNet is a lexical database built by using the same method in BabelNet. We attempt to map the Wikipedia page to WordNet database but only focus on the word related to the domain of business. Also, to enrich the vocabulary in the business domain, we also use terms in the business-specific online dictionary (businessdictionary.com). We utilize this database to do word sense disambiguation process on business process model activity’s terms. The result from this study shows that the database can increase the accuracy of the word sense disambiguation process especially in particular terms related to the business and industrial domains.
Public textual cyberbullying has become one of the most prevalent issues associated with online safety of young people, particularly on social networks. To address this issue, we argue that the boundaries of what constitutes public textual cyberbullying needs to be first identified and a corresponding linguistically motivated definition needs to be advanced. Thus, we propose a definition of public textual cyberbullying that contains three necessary and sufficient elements: the personal marker, the dysphemistic element and the cyberbullying link between the previous two elements. Subsequently, we argue that one of the cornerstones in the overall process of mitigating the effects of cyberbullying is the design of a cyberbullying lexical database that specifies what linguistic and cyberbullying specific information is relevant to the detection process. In this vein, we propose a novel cyberbullying lexical database based on the definition of public textual cyberbullying. The overall architecture of our cyberbullying lexical database is determined semantically, and, in order to facilitate cyberbullying detection, the lexical entry encapsulates two new semantic dimensions that are derived from our definition: cyberbullying function and cyberbullying referential domain. In addition, the lexical entry encapsulates other semantic and syntactic
We present a method for automatically converting the Dutch Lassy Small treebank, a phrasal dependency treebank, to UD. All of the information required to produce accurate UD annotation appears to be available in the underlying annotation. However, we also note that the close connection between POS-tags and dependency labels that is present in UD is missing in the Lassy treebanks. As a consequence, annotation decisions in the Dutch<br/>data for such phenomena as nominalization<br/>and clausal complements of prepositions<br/>seem to differ to some extent from comparable data in English and German. Because the conversion is automatic, we can now also compare three state-of-theart dependency parsers trained on UD Lassy Small with Alpino, a hybrid Dutch parser which produces output that is compatible with the original Lassy annotations.
Svetlana Toldova, Dina Pisarevskaya, Margarita Ananyeva, Maria Kobozeva, Alexander Nasedkin, Sofia Nikiforova, Irina Pavlova, Alexey Shelepov. Proceedings of the 6th Workshop on Recent Advances in RST and Related Formalisms. 2017.
Discourse parsing has long been treated as a stand-alone problem independent from constituency or dependency parsing. Most attempts at this problem are pipelined rather than end-to-end, sophisticated, and not self-contained: they assume goldstandard text segmentations (Elementary Discourse Units), and use external parsers for syntactic features.
This paper introduces the Universal Dependencies Treebank for Slovenian. We overview the existing dependency treebanks for Slovenian and then detail the conversion of the ssj200k treebank to the framework of Universal Dependencies version 2. We explain the mapping of part-of-speech categories, morphosyntactic features, and the dependency relations, focusing on the more problematic language-specific issues. We conclude with a quantitative overview of the treebank and directions for further work.
OBJECTIVE: Intrusive negative affect and concurrent deficits in positive affect are hallmarks of posttraumatic stress disorder (PTSD). We sought to further extend the extant literature by exploring the experience of negative affect intrusion upon potentially positive situations (here termed, "negative affect interference," NAI). METHOD: Two studies with adults endorsing at least 1 traumatic event (Study 1, N = 294; Study 2, N = 286) examined how NAI and more general hedonic deficits (HD) relate to psychopathology, trauma exposure characteristics, and ratings of normed visual stimuli. RESULTS: Study 1 found that NAI and HD were positively correlated with PTSD symptoms and childhood trauma, and NAI incremented over depressive symptoms in predicting PTSD severity. Study 2 results indicated additional strong positive correlations between NAI and HD and anhedonia, affect regulation problems, negative affect, and neuroticism. NAI and HD were found to increment over trait NA in predicting PTSD symptoms. Individuals endorsing elevated NAI and HD rated positively valenced pictures (including food and erotic images) as less arousing, although not more negative. CONCLUSIONS: These findings expand conceptualizations of anhedonia and emotional numbing by drawing attention to negative affect in otherwise positive contexts. (PsycINFO Database Record
Accurate sentiment analysis models encode the sentiment of words and their combinations to predict the overall sentiment of a sentence. This task becomes challenging when applied to morphologically rich languages (MRL). In this article, we evaluate the use of deep learning advances, namely the Recursive Neural Tensor Networks (RNTN), for sentiment analysis in Arabic as a case study of MRLs. While Arabic may not be considered the only representative of all MRLs, the challenges faced and proposed solutions in Arabic are common to many other MRLs. We identify, illustrate, and address MRL-related challenges and show how RNTN is affected by the morphological richness and orthographic ambiguity of the Arabic language. To address the challenges with sentiment extraction from text in MRL, we propose to explore different orthographic features as well as different morphological features at multiple levels of abstraction ranging from raw words to roots. A key requirement for RNTN is the availability of a sentiment treebank; a collection of syntactic parse trees annotated for sentiment at all levels of constituency and that currently only exists in English. Therefore, our contribution also includes the creation of the first Arabic Sentiment Treebank (A r S en TB) that is morphologically and orthographically enriched. Experimental results show that, compared to the basic RNTN proposed for English, our solution achieves significant improvements up to 8% absolute at the phrase level and 10.8% absolute at the sentence level, measured by average F1 score. It also outperforms well-known classifiers including Support Vector Machines, Recursive Auto Encoders, and Long Short-Term Memory by 7.6%, 3.2%, and 1.6% absolute respectively, all models being trained with similar morphological considerations.