_id stringlengths 4 10 | text stringlengths 0 18.4k | title stringlengths 0 8.56k |
|---|---|---|
d621025 | We describe a method for analysing the temporal structure of a discourse which takes into account the effects of tense, aspect, temporal adverbials and rhetorical structure and which minimises unnecessary ambiguity in the temporal structure. It is part of a discourse grammar implemented in Carpenter's ale formalism. The method for building up the temporal structure of the discourse combines constraints and preferences: we use constraints to reduce the number of possible structures, exploiting the hpsg type hierarchy and unification for this purpose; and we apply preferences to choose between the remaining options using a temporal centering mechanism. We end by recommending that an underspecified representation of the structure using these techniques be used to avoid generating the temporal/rhetorical structure until higher-level information can be used to disambiguate. | Algorithms for Analysing the Temporal Structure of Discourse * † |
d9839359 | Probabilistic Models of Grammar Acquisition | |
d227231527 | To explore the potential sembanking in Korean and ways to represent the meaning of Korean sentences, this paper reports on the process of applying Abstract Meaning Representation to Korean, a semantic representation framework that has been studied in a wide range of languages, and its output: the Korean AMR corpus. The corpus which is constructed so far is a size of 1,253 sentences and its raw texts are from ExoBrain Corpus, a state-led R&D project on language AI. This paper also analyzes the result in both qualitative and quantitative manners, proposing discussions for further development.Related WorksEver since AMR was first proposed inBanarescu et al. (2013), efforts to construct corpus have continued in English-speaking nations. The Little Prince Corpus, containing 1,562 sentences, has been continuously providing the foundation for multilingual AMR research; Bio AMR Corpus, comprising 6,952 sentences, is known for proving its applicability in the biomedical domain. 1 In particular, the release of Abstract Meaning Representation (AMR) Annotation Release 3.0 (Knight et al., 2020) demonstrates how the English AMR managed to enter a stable phase.AMR research in non-English-speaking nations has begun its expansion; recent years have seen concrete developments in corpus construction in various parts of the world. In 2014, 100-sentence-sized Chinese and Czech AMR corpus was first built for basic research on multilingual AMR annotation. This work is licensed under a Creative Commons Attribution 4.0 International License. License details: | Building Korean Abstract Meaning Representation Corpus |
d2645484 | Feature structures are a representational device used in several current linguistic theories. This paper shows how these structures can be axiomatized in a decidable class of first-order logic, which can also be used to express constraints on these structures. Desirable properties, such as compactness and decidability, follow directly. Moreover, additional types of feature values, such as "set-valued" features, can be incorporated into the system simply by axiomatizing their properties. | Features and Formulae |
d25492437 | We present a successfully implemented document repository REST service for flexible SCRUD (search, create, read, update, delete) storage of social media and speech conversations, using a GATE/TIPSTER-like document object model and providing a query language for document features. This software is currently being used in the SENSEI research project and will be published as open-source software before the project ends. It is, to the best of our knowledge, the first freely available, general purpose data repository to support large-scale multimodal (i.e., speech or text) conversation analytics. | A Document Repository for Social Media and Speech Conversations |
d1648359 | This paper proposes two approaches to extract translation term pairs from Chinese-English bilingual corpus with more than 500,000 patents. One approach is precision-oriented, in which we compare six term alignment methods. Based on our experiments, we find that the EM (Expectation Maximization) method is the best. However, it is time-consuming and hard to extract many-to-many translations for the same concept. While the MI (mutual information) method performs worst, the term pairs extracted may be totally different from those by EM. This may inspire subsequent researches to study the possibility of hybrid term alignment methods. The other approach is recall-oriented, in which a simple idea was proposed. With an efficient implementation, 20% more term pairs were extracted based on an existing lingual lexicon which already has more than one million term pairs merged from several sources. | Automatic Term Pair Extraction from Bilingual Patent Corpus |
d30799516 | tiden me d at anvende logikprogrammering som hjaelpemiddel ved automa tiseret oversaettelse fra japansk til engelsk og dansk. Logikprogram m e r i n g s s p r o g e t P r ol og a n v e n d e s t i l s y n t a k s a n a l y s e af j a p a n s k , opbygning af en praedikatlogisk repraesentation af teksten, fortolkning af de n n e og ved h j ae l p af o r d b ø g e r o v e r s ae t t e l s e til e n g e l s k eller dansk.Bemaerkninger om det japanske sprog.Japansk er ikke b e s l ae g t e t m e d n o g e t a n d e t sp r o g (måske bortset fra koreansk), og det har sin egen måde at beskrive grammatik på.Indo-europaeiske begreber passer i virkeligheden ikke saerligt godt til japansk. På trods af d e t t e vil b e s k r i v e l s e n af d e t j a p a n s k e s p r o g an vende i n d o -e u r o p ae i s k t e r m i n o l o g i , da det formodes, at laeseren er mest fortrolig m e d denne. Der er a l t s å tale o m et v i s t misforhold mellem beskrive-måde og det beskrevne. Endvidere er der af f r e m s t i ll i n g s m ae s s s i g e g r u n d e f o r e t a g e t n o g l e s i m p l i f i c e r i n g e r , f.eks. i omtalen af partikler. 1.1 Saetningskonstruktion. Principielt er enhver o r d r ae k k e f ø l g e till ad t, blot verbet kommer til sidst. I de allerfleste tilfaelde vil man dog se denne raekkefølge: subjekt indirekte-objekt objekt verbum.Subjektet udelades som regel, hvis det fremgår af sammenhaengen. Kasus angives ved e f t e r s t i l l e d e p a r t i k l e r , o g s å k a l d e t p o s t p o s i t i o n e r .Blandt de vigtigste partikler kan naevnes: "wa" el. "ga" (subjektspartikel), "ni" (indir. objektspartikel), "o" (objektspartikel). Der er dog en vis forskel på "wa" og "ga", som det vil føre for vidt at komme ind på her. Enhver underordnet s ae t n i n g skal k o m m e før d e n ti lh ør en de hovedsaet ning. Da v e r b e t a l t i d står s i d s t i en s ae t n i n g , v i l v e r b e t i en re lativ s ae tn in g stå u m i d d e l b a r t f o r a n de t s u b s t a n t i v , s o m de t er relativt til. Der findes ikke relative pronominer på japansk. 29 Logik anvendt til oversae ttelse af japansk Arendse Bernth Proceedings of NODALIDA 1983, pages 29-36 Endvidere er det naturligvis essentielt for en g o d o v e r s ae t t e l s e , at systemet har adgang til en beskrivelse af et relevant verdensbillede, hvilket også ville hjaelpe til at fastlaegge meningen. Dette er et ikke uvaesentligt problem, som vi dog ikke beskaeftiger os med.4. Littciratvii:. | |
d5344828 | Assessing the semantic similarity between sentences in different languages is challenging.We approach this problem by leveraging multilingual distributional word representations, where similar words in different languages are close to each other. The availability of parallel data allows us to train such representations on a large amount of languages. This allows us to leverage semantic similarity data for languages for which no such data exists. We train and evaluate on five language pairs, including English, Spanish, and Arabic. We are able to train wellperforming systems for several language pairs, without any labelled data for that language pair. | Cross-lingual Learning of Semantic Textual Similarity with Multilingual Word Representations |
d199379780 | Despite recent advances in the application of deep neural networks to various kinds of medical data, extracting information from unstructured textual sources remains a challenging task. The challenges of training and interpreting document classification models are amplified when dealing with small and highly technical datasets, as are common in the clinical domain. Using a dataset of de-identified clinical letters gathered at a memory clinic, we construct several recurrent neural network models for letter classification, and evaluate them on their ability to build meaningful representations of the documents and predict patients' diagnoses. Additionally, we probe sentence embedding models in order to build a humaninterpretable representation of the neural network's features, using a simple and intuitive technique based on perturbative approaches to sentence importance. In addition to showing which sentences in a document are most informative about the patient's condition, this method reveals the types of sentences that lead the model to make incorrect diagnoses. Furthermore, we identify clusters of sentences in the embedding space that correlate strongly with importance scores for each clinical diagnosis class. | Analysing Representations of Memory Impairment in a Clinical Notes Classification Model |
d252819227 | This paper documents our approach for the Creative-Summ 2022 shared task for Automatic Summarization of Creative Writing. For this purpose, we develop an automatic summarization pipeline where we leverage a denoising autoencoder for pretraining sequence-to-sequence models and fine-tune it on a large-scale abstractive screenplay summarization dataset to summarize TV transcripts from primetime shows. Our pipeline divides the input transcript into smaller conversational blocks, removes redundant text, summarises the conversational blocks, obtains the block-wise summaries, cleans, structures, and then integrates the summaries to create the meeting minutes. Our proposed system achieves first position with some of the best scores across multiple metrics (lexical, semantical) in the Creative-Summ shared task. We publicly release our proposed system here 1 | Automatic Summarization for Creative Writing: Denoising Auto-Encoder based Pipeline Method for Generating Summary of Movie Scripts |
d878415 | T1LANSFORMATIONAL G1LAMMAI:k AND TKANSFORMA- TIONAL PARSING IN THE REQUEST SYSTEM | |
d3893109 | On Formal Versus Commonsense Semantics | |
d5668932 | This paper describes a mechanism for identifying errors made by a student during a computer-aided language learning dialogue. The mechanism generates a set of 'perturbations' of the student's original typed utterance, each of which embodies a hypothesis about an error made by the student. Perturbations are then passed through the system's ordinary utterance interpretation pipeline, along with the student's original utterance. An utterance disambiguation algorithm selects the best interpretation, performing error correction as a side-effect. | Error correction using utterance disambiguation techniques |
d65130 | ||
d23646123 | Wordnets are rich lexico-semantic resources.Linked wordnets are extensions of wordnets, which link similar concepts in wordnets of different languages. Such resources are extremely useful in many Natural Language Processing (NLP) applications, primarily those based on knowledge-based approaches. In such approaches, these resources are considered as gold standard/oracle. Thus, it is crucial that these resources hold correct information. Thereby, they are created by human experts. However, manual maintenance of such resources is a tedious and costly affair. Thus techniques that can aid the experts are desirable. In this paper, we propose an approach to link wordnets. Given a synset of the source language, the approach returns a ranked list of potential candidate synsets in the target language from which the human expert can choose the correct one(s). Our technique is able to retrieve a winner synset in the top 10 ranked list for 60% of all synsets and 70% of noun synsets. | Semi-automatic WordNet Linking using Word Embeddings |
d752448 | This paper presents the task definition, resources, participating systems, and comparative results for a Romanian Word Sense Disambiguation task, which was organized as part of the SENSEVAL-3 evaluation exercise. Five teams with a total of seven systems were drawn to this task. | An Evaluation Exercise for Romanian Word Sense Disambiguation |
d202634572 | 1The present study has surveyed professional translators working in six international organizations in order to know more about their views and attitudes with regard to new translation workflows involving two different types of technologies, i.e. machine translation and speech recognition. The main aim of this survey was to identify how feasible it is to implement new post-editing workflows in an international organization using speech as an input method to edit inaccurate machine translation outputs. Overall, the results suggest that the surveyed translators do not hold a negative view on the use of ASR as part of their translation workflow, which provides a promising first step towards investigating the integration of speech based post-editing to translation workflows for productivity and ergonomic gains. | Surveying the potential of using speech technologies for post-editing purposes in the context of international organizations: What do p rofessional translators think? |
d248780536 | Over the years, there has been a slow but steady change in the attitude of society towards different kinds of sexuality. However, on social media platforms, where people have the license to be anonymous, toxic comments targeted at homosexuals, transgenders and the LGBTQ+ community are not uncommon. Detection of homophobic comments on social media can be useful in making the internet a safer place for everyone. For this task, we used a combination of word embeddings and SVM Classifiers as well as some BERT-based transformers. We achieved a weighted F1-score of 0.93 on the English dataset, 0.75 on the Tamil dataset and 0.87 on the Tamil-English Code-Mixed dataset. | SSNCSE_NLP@LT-EDI-ACL2022: Homophobia/Transphobia Detection in Multiple Languages using SVM Classifiers and BERT-based Transformers |
d202764963 | ||
d17652442 | The biomedical domain offers a wealth of linguistic resources for Natural Language Processing, including terminologies and corpora. While many of these resources are prominently available for English, other languages including French benefit from substantial coverage thanks to the contribution of an active community over the past decades. However, access to terminological resources in languages other than English may not be as straight-forward as access to their English counterparts. Herein, we review the extent of resource coverage for French and give pointers to access French-language resources. We also discuss the sources and methods for making additional material available for French. | Language Resources for French in the Biomedical Domain |
d5152228 | A method of sense resolution is proposed that is based on WordNet, an on-line lexical database that incorporates semantic relations (synonymy, antonymy, hyponymy, meronymy, causal and troponymic entailment) as labeled pointers between word senses. With WordNet, it is easy to retrieve sets of semantically rehted words, a facility that will be used for sense resolution during text processing, as follows. When a word with multiple senses is encountered, one of two procedures will be followed. Either, (1) words related in meaning to the alternative senses of the polysemous word will be retrieved; new strings will be derived by substituting these related words into the context of the polysemous word; a large textual corpus will then be searched for these derived strings; and that sense will be chosen that corresponds to the derived string that is found most often in the corpus. Or, (2) the context of the polysemous word will be used as a key to search a large corpus; all words found to occur in that context will be neted; Word.Net will then be used to estimate the semantic distance from those words to the alternative senses of the polysemous word; and that sense will be chosen that is closest in meaning to other words occurring in the same contexL If successful, this procedure could have practical applications to problems of information retrieval, mechan/cal translation, intelligent tutoring systems, and elsewhere.BACKGROUNDAn example can set the problem. Suppose that an automatic transcription device were to recognize the string of phonemes/ralt/in the flow of speech and could correctly identify it as an English word; the device would still have to decide whether the word should be spelled right, write, or rite. And if the result were then | A PROPOSAL FOR LEXICAL DISAMBIGUATION |
d236486168 | ||
d252819390 | Humans use different wordings depending on the context to facilitate efficient communication. For example, instead of completely new information, information related to the preceding context is typically placed at the sentence-initial position. In this study, we analyze whether neural language models (LMs) can capture such discourse-level preferences in text generation. Specifically, we focus on a particular aspect of discourse, namely the topic-comment structure. To analyze the linguistic knowledge of LMs separately, we chose the Japanese language, a topicprominent language, for designing probing tasks, and we created human topicalization judgment data by crowdsourcing. Our experimental results suggest that LMs have different generalizations from humans; LMs exhibited less context-dependent behaviors toward topicalization judgment. These results highlight the need for the additional inductive biases to guide LMs to achieve successful discourselevel generalization. | Topicalization in Language Models: A Case Study on Japanese |
d5231912 | Most previous work in information extraction from text has focused on named-entity recognition, entity linking, and relation extraction. Less attention has been paid given to extracting the temporal scope for relations between named entities; for example, the relation president-Of( | Temporal Scoping of Relational Facts based on Wikipedia Data |
d10916069 | WordNet::SenseRelate::AllWords is a freely available open source Perl package that assigns a sense to every content word (known to WordNet) in a text. It finds the sense of each word that is most related to the senses of surrounding words, based on measures found in WordNet::Similarity. This method is shown to be competitive with results from recent evaluations including SENSEVAL-2 and SENSEVAL-3. | WordNet::SenseRelate::AllWords - A Broad Coverage Word Sense Tagger that Maximizes Semantic Relatedness |
d15406016 | We introduce a method for transferring annotation from a syntactically annotated corpus in a source language to a target language. Our approach assumes only that an (unannotated) text corpus exists for the target language, and does not require that the parameters of the mapping between the two languages are known. We outline a general probabilistic approach based on Data Augmentation, discuss the algorithmic challenges, and present a novel algorithm for sampling from a posterior distribution over trees. | Treebank Transfer |
d9740006 | This paper investigates the use of sentential pronouns in English and Norwegian. We argue that resolution of sentential pronouns is sensitive to the distinction between forms whose referents must be in focus and forms whose referents must only be activated, but not necessarily in focus. An investigation of the distribution and interpretation of sentential pronouns also reveals that the relative salience of a higher order discourse endty is influenced by syntactic structure as well as extralinguistic factors. | |
d39590661 | Japanese coordinate noun phrases by the particle TO are often ambiguous on whether it means two parallel propositions (and) ,or a mutual case relation (with or against) , as deep case structure. It was a hard problem to determine it, though they are widely used. We propose a method of solving the ambiguity by analyzing mutualness of verbs and adjectives. The mutualness is determined by three features of each verb or adjective. The first feature indicates permission of mutual expression in the subject, and the second in the object. The last shows if a verb is voluntary.Using this method we design a parsing mechanism, where matching of features is represented as neutralization between predicate arguments. | Disambiguation of Coordinate Expressions in Japanese by Extracting Mutual Case Relation |
d10080734 | Languages are constantly evolving through their users due to the need to communicate more efficiently. Under this hypothesis, we formulate unsupervised word segmentation as a regularized compression process. We reduce this process to an optimization problem, and propose a greedy inclusion solution. Preliminary test results on the Bernstein-Ratner corpus and Bakeoff-2005 show that the our method is comparable to the state-of-the-art in terms of effectiveness and efficiency. | A Regularized Compression Method To Unsupervised Word Segmentation |
d201762525 | Torwali is an endangered language spoken in the north of Pakistan. It is a computationally challenging language because of its RTL Perso-Arabic script, non-concatenative nature and distinct words alterations. This paper discusses issues and challenges regarding grammatical structure, divergence in terms of lexicon as well as morphological makeup for the machine translation of a less studied language. It includes creation of NLP tools such as parts of speech (POS) tagger and morphological analyser with HFST which is based on the idea of building lexicon and morphological rules using finite state devices. This work, on which this paper is based, will be a source of Torwali finite state morphology and its future computational growth as electronic dictionaries are usually equipped with morphological analyser and it will also be helpful for developing language pairs. | A step towards Torwali machine translation: an analysis of morphosyntactic challenges in a low-resource language |
d16248625 | We show that asymmetric models based onTversky (1977)improve correlations with human similarity judgments and nearest neighbor discovery for both frequent and middle-rank words. In accord with Tversky's discovery that asymmetric similarity judgments arise when comparing sparse and rich representations, improvement on our two tasks can be traced to heavily weighting the feature bias toward the rarer word when comparing high-and midfrequency words. | Improving sparse word similarity models with asymmetric measures |
d252819353 | Question generation is the task of automatically generating questions based on given context and answers, and there are problems that the types of questions and answers do not match. In minority languages such as Tibetan, since the grammar rules are complex and the training data is small, the related research on question generation is still in its infancy. To solve the above problems, this paper constructs a question type classifier and a question generator. We perform fine-grained division of question types and integrate grammatical knowledge into question type classifiers to improve the accuracy of question types. Then, the types predicted by the question type classifier are fed into the question generator. Our model improves the accuracy of interrogative words in generated questions, and the BLEU-4 on SQuAD reaches 17.52, the BLEU-4 on HotpotQA reaches 19.31, the BLEU-4 on TibetanQA reaches 25.58. | Question Generation Based on Grammar Knowledge and Fine-grained Classification |
d16012692 | Semantic text processing faces the challenge of defining the relation between lexical expressions and the world to which they make reference within a period of time. It is unclear whether the current test sets used to evaluate disambiguation tasks are representative for the full complexity considering this time-anchored relation, resulting in semantic overfitting to a specific period and the frequent phenomena within. We conceptualize and formalize a set of metrics which evaluate this complexity of datasets. We provide evidence for their applicability on five different disambiguation tasks. To challenge semantic overfitting of disambiguation systems, we propose a time-based, metric-aware method for developing datasets in a systematic and semi-automated manner, as well as an event-based QA task. | Semantic overfitting: what 'world' do we consider when evaluating disambiguation of text? |
d51918720 | In this paper, we apply the contribution model of grounding to a corpus of humanhuman peer-mentoring dialogues. From this analysis, we propose effective turntaking strategies for human-robot interaction with a teachable robot. Specifically, we focus on (1) how robots can encourage humans to present and (2) how robots can signal that they are going to begin a new presentation. We evaluate the strategies against a corpus of human-robot dialogues and offer three guidelines for teachable robots to follow to achieve more humanlike collaborative dialogue. | Turn-Taking Strategies for Human-Robot Peer-Learning Dialogue |
d6860686 | This paper explores the possibility of using the paradigm of Dynamic Logic (DL) to formalise information states and update processes on information states. In particular, we present a formalisation of the dialogue gameboard introduced by Jonathan Ginzburg. From a more general point of view, we show that DL is particularly well suited to develop rigorous formal foundations for an approach to dialogue dynamics based on information state updates. | A Dynamic Logic Formalisation of the Dialogue Gameboard |
d221819674 | We propose AutoQA, a methodology and toolkit to generate semantic parsers that answer questions on databases, with no manual effort. Given a database schema and its data, AutoQA automatically generates a large set of high-quality questions for training that covers different database operations. It uses automatic paraphrasing combined with templatebased parsing to find alternative expressions of an attribute in different parts of speech. It also uses a novel filtered auto-paraphraser to generate correct paraphrases of entire sentences.We apply AutoQA to the Schema2QA dataset and obtain an average logical form accuracy of 62.9% when tested on natural questions, which is only 6.4% lower than a model trained with expert natural language annotations and paraphrase data collected from crowdworkers. To demonstrate the generality of AutoQA, we also apply it to the Overnight dataset. AutoQA achieves 69.8% answer accuracy, 16.4% higher than the state-of-the-art zero-shot models and only 5.2% lower than the same model trained with human data. | AutoQA: From Databases To QA Semantic Parsers With Only Synthetic Training Data |
d2330504 | We present a comparison between two systems for establishing syntactic and semantic dependencies: one that performs dependency parsing and semantic role labeling as a single task, and another that performs the two tasks in isolation. The systems are based on local memorybased classifiers predicting syntactic and semantic dependency relations between pairs of words. In a second global phase, the systems perform a deterministic ranking procedure in which the output of the local classifiers is combined per sentence into a dependency graph and semantic role labeling assignments for all predicates. The comparison shows that in the learning phase a joint approach produces better-scoring classifiers, while after the ranking phase the isolated approach produces the most accurate syntactic dependencies, while the joint approach yields the most accurate semantic role assignments. | Dependency Parsing and Semantic Role Labeling as a Single Task |
d16848750 | Source-side reordering has recently seen a surge in popularity in machine translation research, often providing enormous reductions in translation time and showing good empirical results in translation quality. For many language pairs, however-especially for translation into morphologically rich languages-the assumptions of these models may be too crude. But while such language pairs call for more complex models, these could increase the search space to an extent that would diminish their benefits. In this paper, we examine the question whether purely syntaxoriented adaptation models (i.e., models only considering word order) can be used as a means to delimit the search space for more complex morphosyntactic models. We propose a model based on a popular preordering algorithm(Lerner and Petrov, 2013). This novel preordering model is able to produce both n-best word order predictions as well as distributions over possible word order choices in the form of a lattice and is therefore a good fit for use by richer models taking into account aspects of both syntax and morphology. We show that the integration of non-local language model features can be beneficial for the model's preordering quality and evaluate the space of potential word order choices the model produces.This work is licenced under a Creative Commons Attribution 4.0 International License. | Delimiting Morphosyntactic Search Space with Source-Side Reordering Models |
d51729732 | Because obtaining training data is often the most difficult part of an NLP or ML project, we develop methods for predicting how much data is required to achieve a desired test accuracy by extrapolating results from systems trained on a small pilot training dataset. We model how accuracy varies as a function of training size on subsets of the pilot data, and use that model to predict how much training data would be required to achieve the desired accuracy. We introduce a new performance extrapolation task to evaluate how well different extrapolations predict system accuracy on larger training sets. We show that details of hyperparameter optimisation and the extrapolation models can have dramatic effects in a document classification task. We believe this is an important first step in developing methods for estimating the resources required to meet specific engineering performance targets. | Predicting accuracy on large datasets from smaller pilot data |
d1801462 | Automatic story generation is the subject of a growing research effort which has mainly focused on fictional stories. In this paper, we present some preliminary work to generate récits (stories) from sensors data acquired during a ski sortie. In this approach, the story planning is performed using a task model that represents domain knowledge and sequential constraints between ski activities. To test the validity of the task model, a small-scale user evaluation was performed to compare the human perception of récit plans from hand written or automatically generated récits. This evaluation showed no difference in story plan identification adding credence to the eligibility of the task model for representing story plan in NLG. To go a step further, a basic NLG system to generate narrative from activities extracted from GPS data is also reported. | Generating récit from sensor data: evaluation of a task model for story planning and preliminary experiments with GPS data |
d2357255 | In recent years microblogs have taken on an important role in the marketing sphere, in which they have been used for sharing opinions and/or experiences about a product or service. Companies and researchers have become interested in analysing the content generated over the most popular of these, the Twitter platform, to harvest information critical for their online reputation management (ORM). Critical to this task is the efficient and accurate identification of tweets which refer to a company distinguishing them from those which do not. The aim of this work is to present and compare two different approaches to achieve this. The obtained results are promising while at the same time highlighting the difficulty of this task. | On the Difficulty of Clustering Microblog Texts for Online Reputation Management |
d17007944 | Multimodal grammars provide an effective mechanism for quickly creating integration and understanding capabilities for interactive systems supporting simultaneous use of multiple input modalities. However, like other approaches based on hand-crafted grammars, multimodal grammars can be brittle with respect to unexpected, erroneous, or disfluent input. In this article, we show how the finite-state approach to multimodal language processing can be extended to support multimodal applications combining speech with complex freehand pen input, and evaluate the approach in the context of a multimodal conversational system (MATCH). We explore a range of different techniques for improving the robustness of multimodal integration and understanding. These include techniques for building effective language models for speech recognition when little or no multimodal training data is available, and techniques for robust multimodal understanding that draw on classification, machine translation, and sequence edit methods. We also explore the use of edit-based methods to overcome mismatches between the gesture stream and the speech stream. * | Robust Understanding in Multimodal Interfaces |
d550145 | Unsupervised learning of grammar is a problem that can be important in many areas ranging from text preprocessing for information retrieval and classification to machine translation. We describe an MDL based grammar of a language that contains morphology and lexical categories. We use an unsupervised learner of morphology to bootstrap the acquisition of lexical categories and use these two learning processes iteratively to help and constrain each other. To be able to do so, we need to make our existing morphological analysis less fine grained. We present an algorithm for collapsing morphological classes (signatures) by using syntactic context. Our experiments demonstrate that this collapse preserves the relation between morphology and lexical categories within new signatures, and thereby minimizes the description length of the model. | Using Morphology and Syntax Together in Unsupervised Learning |
d251403821 | This paper examines the state of data protection and privacy in the United States. There is no comprehensive federal data protection or data privacy law despite bipartisan and popular support. There are several data protection bills pending in the 2022 session of the US Congress, five of which are examined in Section 2 below. Although it is not likely that any will be enacted, the growing number reflects the concerns of citizens and lawmakers about the power of big data. Recent actions against data abuses, including data breaches, litigation and settlements, are reviewed in Section 3 of this paper. These reflect the real harm caused when personal data is misused. Section 4 contains a brief US copyright law update on the fair use exemption, highlighting a recent court decision and indications of a re-thinking of the fair use analysis. In Section 5, some observations are made on the role of privacy in data protection regulation. It is argued that privacy should be considered from the start of the data collection and technology development process. Enhanced awareness of ethical issues, including privacy, through university-level data science programs will also lay the groundwork for best practices throughout the data and development cycles. | Data Protection, Privacy and US Regulation |
d9825180 | We present a simple yet effective approach to syntactic reordering for Statistical Machine Translation (SMT). Instead of solely relying on the top-1 best-matching rule for source sentence preordering, we generalize fully lexicalized rules into partially lexicalized and unlexicalized rules to broaden the rule coverage. Furthermore, , we consider multiple permutations of all the matching rules, and select the final reordering path based on the weighed sum of reordering probabilities of these rules.Our experiments in English-Chinese and English-Japanese translations demonstrate the effectiveness of the proposed approach: we observe consistent and significant improvement in translation quality across multiple test sets in both language pairs judged by both humans and automatic metric. | Generalized Reordering Rules for Improved SMT |
d7015966 | This work is supported by an IRCSET Ph.D. Fellowship Award. | Hybridity in MT: Experiments on the Europarl Corpus |
d15273963 | This paper is concerned with a particular kind of Korean relative clause constructions that contain possessive specifier gaps, as opposed to complement or adjunct gaps. Those relative clauses have been known to violate Ross's Complex NP Constraint, and various attempts have been made to account for the apparent island violations. This paper argues that those relative clauses have possessive specifier gaps in the main relative clause, rather than in the embedded relative clause, and so there are no island violations involved in such relative clauses. It is argued that the island violations in question can be viewed as violations of selection restrictions of very general kind holding between nouns and their specifers. It is not necessary to impose any special constraints to account for the apparent violations observed in the particular kind of relative clauses discussed in this paper. | RELATIVE CLAUSE CONSTRUCTIONS WITH POSSESSIVE SPECIFIER GAPS: A CONSTRAINT-BASED APPROACH * |
d2002468 | A pseudoword is a composite comprised of two or more words chosen at random; the individual occurrences of the original words within a text are replaced by their conflation. Pseudowords are a useful mechanism for evaluating the impact of word sense ambiguity in many NLP applications. However, the standard method for constructing pseudowords has some drawbacks. Because the constituent words are chosen at random, the word contexts that surround pseudowords do not necessarily reflect the contexts that real ambiguous words occur in. This in turn leads to an optimistic upper bound on algorithm performance. To address these drawbacks, we propose the use of lexical categories to create more realistic pseudowords, and evaluate the results of different variations of this idea against the standard approach. | Category-Based Pseudowords |
d252624580 | Like most other minority languages, Scottish Gaelic has limited tools and resources available for Natural Language Processing research and applications. These limitations restrict the potential of the language to participate in modern speech technology, while also restricting research in fields such as corpus linguistics and the Digital Humanities. At the same time, Gaelic has a long written history, is well-described linguistically, and is unusually well-supported in terms of potential NLP training data. For instance, archives such as the School of Scottish Studies hold thousands of digitised recordings of vernacular speech, many of which have been transcribed as paper-based, handwritten manuscripts. In this paper, we describe a project to digitise and recognise a corpus of handwritten narrative transcriptions, with the intention of re-purposing it to develop a Gaelic speech recognition system. | Handwriting Recognition for Scottish Gaelic |
d221373775 | Dans cet article, nous présentons la mise en oeuvre d'une chaîne de traitement sémantique complète dédiée aux conversations audio issues de centres d'appel téléphoniques, depuis la phase de transcription automatique jusqu'à l'exploitation des résultats, en passant par l'étape d'analyse sémantique des énoncés. Nous décrivons ici le fonctionnement des différentes analyses que notre équipe développe, ainsi que la plateforme interactive permettant de restituer les résultats agrégés de toutes les conversations analysées.ABSTRACTSemantic analysis of automatic phone call transcriptions in FrenchIn this article, we present the implementation of a complete semantic processing chain dedicated to call center phone conversations, from the speech-to-text phase to the exploitation of the results, including the semantic analysis of the utterances. Here we describe the workings of the various analyses that our team develops, as well as the interactive platform that displays the aggregated results of all the analyzed conversations. | Analyse sémantique de transcriptions automatiques d'appels téléphoniques en français |
d5235868 | We study the interplay of the discourse structure of a scientific argument with formal citations. One subproblem of this is to classify academic citations in scientific articles according to their rhetorical function, e.g., as a rival approach, as a part of the solution, or as a flawed approach that justifies the current research. Here, we introduce our annotation scheme with 12 categories, and present an agreement study. | An annotation scheme for citation function |
d5440348 | In this paper a system which understands and conceptualizes scenes descriptions in natural language is presented. Specifically, the following components of the system are described: the syntactic analyzer, based on a Procedural Systemic Grammar, the semantic analyzer relying on the Conceptual Dependency Theory, and the dictionary. I | NATURAL LANGUAGE INPUT FOR SCENE GENERATION M |
d250390960 | This paper presents a corpus of 43,985 clinical patient notes (PNs) written by 35,156 examinees during the high-stakes USMLE ® Step 2 Clinical Skills examination. In this exam, examinees interact with standardized patientspeople trained to portray simulated scenarios called clinical cases. For each encounter, an examinee writes a PN, which is then scored by physician raters using a rubric of clinical concepts, expressions of which should be present in the PN. The corpus features PNs from 10 clinical cases, as well as the clinical concepts from the case rubrics. A subset of 2,840 PNs were annotated by 10 physician experts such that all 143 concepts from the case rubrics (e.g., shortness of breath) were mapped to 34,660 PN phrases (e.g., dyspnea, difficulty breathing). The corpus is available via a data sharing agreement with NBME and can be requested at https://www.nbme.org/ services/data-sharing. | The USMLE ® Step 2 Clinical Skills Patient Note Corpus |
d243864610 | What is the best way to learn embeddings for entities, and what can be learned from them? We consider this question for the case of literary characters. We address the highly challenging task of guessing, from a sentence in the novel, which character is being talked about, and we probe the embeddings to see what information they encode about their literary characters. We find that when continuously trained, entity embeddings do well at the masked entity prediction task, and that they encode considerable information about the traits and characteristics of the entities. | "Politeness, you simpleton!" retorted [MASK]: Masked prediction of literary characters |
d61280852 | 0.Introduction 0.1. The last time I had the privilege of addressing the Aslib technical translation group, some fifteen years ago, I described my concept of training for specialised translators. At that time the newer universities were developing new language degree courses as an alternative to the traditional language and literature courses, the then Federation of British Industries organised symposia on the need for languages in industry and the Government began seriously to look towards Europe. Since then the country has become more language conscious -as can be seen by the greater volume of translations required, the corresponding increase in the number of professional translators, the diverse employment opportunities for graduates with knowledge of a foreign language and indeed the greater contacts with Europe at all levels. The task of translation itself has remained the same, it still is what it has always been, a mediating function variously called a skill or an art requiring at the same time creativity, tact, and a self-effacing nature. It is routine and tedium to some and an intellectual challenge to others. What has changed, however, are the tools of the trade or profession and this seminar provides an opportunity to survey the tools now available, and to assess their achievements and potential.0.2.Our topic is translation and the computer, which to many people, especially translators, represents an unholy alliance or indeed a disjunction rather than a conjunction. There is no need for translators to consider becoming Luddites. | MULTILINGUAL COMMUNICATION: CHAIRMAN'S INTRODUCTORY REVIEW OF TRANSLATING AND THE COMPUTER The processes of producing and understanding language together with the process of translating messages from one language |
d38027754 | ____________________________________________________________________________________________________________Cet article aborde la question suivante : peut-on mettre en évidence un transfert prosodique du corse (une langue italo-romane) vers le français parlé en Corse, où le français est maintenant la langue dominante ? Un corpus de phrases transparentes en corse et en français telles que a turista trova a caserna (« la touriste trouve la caserne ») a été mis au point, et les productions de locuteurs bilingues enregistrés en Corse ont été comparées avec les contreparties françaises de locuteurs parisiens de référence. Il apparaît que la mélodie des questions totales différencie d'un côté le corse et le français de Corse (avec tous deux des tons hauts suivis de descentes mélodiques finales), de l'autre le français standard (avec des tons hauts en fin de question). Ce premier patron peut être interprété comme un transfert prosodique du corse vers le français.ABSTRACT _________________________________________________________________________________________________________Corsican questions: is there a prosodic transfer from Corsican to French?This study investigates whether a prosodic transfer can be highlighted from Corsican (an Italo-Romance language) to French spoken in Corsica, where French is now the dominant language. A corpus of transparent sentences such as la touriste trouve la caserne (French) or a turista trova a caserna (Corsican) was designed and the productions of bilingual speakers, recorded in Corsica, were compared with the French counterparts of Parisian reference speakers. The melody of yes/no questions turns out to contrast Corsican and Corsican French (both with high tones followed by final pitch falls) and standard French (with utterance-final high tones). The former pattern can be interpreted as a prosodic transfer from Corsican to French. MOTS-CLÉS : prosodie en contact, questions, accent corse en français, langues en danger. | Questions corses : peut-on mettre en évidence un transfert prosodique du corse vers le français ? |
d5814013 | We developed three systems based on automatic paraphrasing techniques to help English learners and English-language beginners. One system extracts personal error patterns in the user's English usage. The second transforms English sentences containing the letters "l" and "r" into sentences containing fewer instances of these letters, which Japanese people have trouble pronouncing properly in English. This system could be used, for example, to transform a draft of a presentation that a Japanese speaker was to present to an audience. The third is an annotation system that provides definition sentences of difficult English words, making them easier to understand. We believe that these systems will be useful both for learners of English and in studies on second-language acquisition. | Three English Learner Assistance Systems Using Automatic Paraphrasing Techniques |
d6358777 | We describe a unified and coherent syntactic framework for supporting a semanticallyinformed syntactic approach to statistical machine translation. Semantically enriched syntactic tags assigned to the target-language training texts improved translation quality. The resulting system significantly outperformed a linguistically naive baseline model (Hiero), and reached the highest scores yet reported on the NIST 2009 Urdu-English translation task. This finding supports the hypothesis (posed by many researchers in the MT community, e.g., in DARPA GALE) that both syntactic and semantic information are critical for improving translation quality-and further demonstrates that large gains can be achieved for low-resource languages with different word order than English. | Semantically-Informed Syntactic Machine Translation: A Tree-Grafting Approach |
d252819002 | Nominal metaphors are frequently used in human language and have been shown to be effective in persuading, expressing emotion, and stimulating interest. This paper tackles the problem of Chinese Nominal Metaphor (NM) generation. We introduce a novel neural framework, which jointly optimizes three tasks: NM identification, NM component identification, and NM generation. The metaphor identification module is able to perform a self-training procedure, which discovers novel metaphors from a large-scale unlabeled corpus for NM generation. The NM component identification module emphasizes components during training and conditions the generation on these NM components for more coherent results. To train the NM identification and component identification modules, we construct an annotated corpus consisting of 6.3k sentences that contain diverse metaphorical patterns. Automatic metrics show that our method can produce diverse metaphors with good readability, where 92% of them are novel metaphorical comparisons. Human evaluation shows our model significantly outperforms baselines on consistency and creativity. * Corresponding author 1. 这个[孩子] tenor 壮的像[牛] vehicle This [boy] tenor is as strong as a [bull] vehicle . Nominal 2. [生活] tenor 好比[旅行] vehicle , 没有计划就难以前行 [Life] tenor is a [journey] vehicle , we cannot move on without a plan. Nominal 3. Meta股价[跳水] metaphorical META stock price [dives] metaphorical . Verbal 4. 他可以像大厨一样烹饪 He can cook like a pro. Literal | CM-Gen: A Neural Framework for Chinese Metaphor Generation with Explicit Context Modelling |
d28720705 | The paper examines how adjectival modification classes can be detected by applying unsupervised methods on adjective-noun co-occurrence data. It evaluates a k-means baseline, two graphical models, and a recently introduced bidirectional clustering algorithm against HeiPLAS, a manually annotated gold standard for hidden modificational classes. The paper shows that the bidirectional clustering algorithm performs best on this task, and discusses how the results of the unsupervised approaches can be employed for building a frame-based inventory of adjectival modification.1 See Anderson and Löbner (2017) for a frame based approach to relational adjectives. | Unsupervised Induction of Compositional Classes for English Adjective-Noun Pairs |
d9200098 | Texts containing personal health information reveal enough data for a third party to be able to identify an individual and his health condition. Detection of personal health information in electronic health records is an essential part of record deidentification. Performance evaluation in use today focuses on method's ability to identify whether a word reveals personal health information or not. In this study, we propose and show that the multi-label classification measures better serve the final goal of the record de-identification. | Evaluation Measures for Detection of Personal Health Information |
d202895775 | ||
d251490836 | This paper describes the conversion of the Sinica Treebank, one of the major Mandarin Chinese treebanks, to Universal Dependencies. The conversion is rule-based and the process involves POS tag mapping, head adjusting in line with the UD scheme and the dependency conversion. Linguistic insights into Mandarin Chinese alongwith the conversion are also discussed. The resulting corpus is the UD Chinese Sinica Treebank which contains more than fifty thousand tree structures according to the UD scheme. The dataset can be downloaded at https://github.com/ckiplab/ud. | Converting the Sinica Treebank of Mandarin Chinese to Universal Dependencies |
d1040974 | Elements of the history, state of the art, and probable future of Machine Translation (MT) are discussed. The treatment is largely tutorial, based on the assumption that this audience is, for the most part, ignorant of matters pertaining to translation in general, and MT in particular. The paper covers some of the major MT R&D groups, the general techniques they employ(ed), and the roles they play(ed) in the development of the field. The conclusions concern the seeming permanence of the translation problem, and potential re-integration of MT with mainstream Computational Linguistics. | A SURVEY OF MACHINE TRANSLATION: ITS HISTORY, CURRENT STATUS, AND FUTURE PROSPECTS |
d7975030 | We report a first major upgrade of Inforex -a web-based system for qualitative and collaborative text corpora annotation and analysis. Inforex is a part of Polish CLARIN infrastructure 1 . It is integrated with a digital repository for storing and publishing language resources 2 and it allows to visualize, browse and annotate text corpora stored in the repository. As a result of a series of workshops for researchers in Humanities and Social Sciences we improved the graphical interface to make the system more friendly and readable for non-experienced users. We also implemented a new functionality for a gold standard annotation which includes private annotations and annotation agreement by a super-annotator. | Inforex -a Collaborative System for Text Corpora Annotation and Analysis |
d10853724 | This paper contains the detailed approach of automatic extraction of Keyphrases from scientific articles (i.e. research paper) using supervised tool like Conditional Random Fields (CRF). | Keyphrase Extraction in Scientific Articles: A Supervised Approach |
d249204438 | This document describes how Yamagata Europe enables organizations to connect seamlessly to its machine translation and translation management system infrastructure using a JSON-based (JavaScript Object Notation) data exchange mechanism. | Connecting client infrastructure with Yamagata Europe machine translation using JSON-based data exchange |
d102745 | 摘要 語意的研究十分依賴語意知識庫所提供的訊息，由於語意研究逐漸變得熱門，相對的語意知 識庫的建構也變得十分迫切。WordNet 是目前最廣為人知的英語語意知識庫，許多語意解歧(word sense disambiguation)的研究都以 WordNet 為共同標準。由於 WordNet 的成功，使得許多其他語 系的 WordNet 建構計畫也紛紛出現。本文提出一個自動從雙語學術名詞庫中抽取中文語意訊息 的方法，這個方法利用一個詞和詞的對應(word-to-word alignment)演算法抽取中英詞對譯的訊 息，再用語意解歧的方法，將中文詞連結到 WordNet synset，以建構中文 WordNet。 1. 緒論 近年來在自然語言處理領域中，語意研究受到了廣泛的重視。語意解歧的技術不斷推陳出 新，進而使得語意的應用也受到鼓舞。然而，語意的使用必需仰賴語意知識庫提供語意訊息，這 些訊息包括一個詞彙有多少不同的語意，以及一個語意和另一個語意是否有同義關係或是上下位 關係等。例如： 「分子」可以表示化學上的「粒子」(如「水分子」)，也可以表示「一群人」(如 「激進分子」)；而「拉布拉多」和「大麥町」由其上位詞可知都是一種「狗」 。 WordNet 是一部訊息豐富的語意知識庫[Miller 1990]，其中收錄了為數極多的詞彙。在結構 上它將所有的相同的語意集成 synset，並以 synset 為基礎進一步連結語意之間的關係，如上位 關係(hypernym)、下位關係(hyponym)、整體關係(holonyms)及部分關係(meronyms)等。目前 WordNet 已經被應用在許多的研究上，如語意解歧(word sense disambiguation)、資訊檢索 (information retrieval) 及電腦輔助語言學習(computer-assisted language learning)等領域，儼然成為 語意研究的共同標準。 由於 WordNet 的成功使得許多其他語系的 WordNet 建構計畫相繼出現。例如： EuroWordNet (EWN)，該計畫目標為建構包含多種歐洲語的 WordNet，及中文詞網計畫[CKIP 2003]，以建構中文語意知識庫為目標。從零開始建構一個 WordNet 是一項艱鉅的任務，所以有 許多的研究嘗試以自動的方式將詞彙連結到 WordNet。例如：[Atserias et al. 1997]、[Daude et al. 1999]以及[Jason et al. 2003]都是利用雙語詞典所提供的翻譯，自動將詞彙連結到 WordNet。使用 一般雙語詞典的翻譯最大的問題在於用詞過度典型化。例如："plant" 在 WordNet 中的第一個語 意 "plant, works, industrial plant"，在雙語詞典中翻譯成「工廠」 。但實際上在文章中可能翻譯成 「廠」 、 「工廠」 、 「廠房」 、 「所」(如「power plant/發電所」)及「工場」等詞。用詞過度典型化的 現象，使得許多文章中的用詞無法找到適當的翻譯連結到 WordNet。 在本實驗中，我們選擇以雙語學術名詞庫作為抽取語意訊息的資料來源。由於學術名詞庫中 包含了大量的複合詞，所以很多詞會搭配不同的詞一再出現，並對應到不同的翻譯。因此不但可 以避免一般雙語詞典翻譯過度典型化的問題，而且多樣化的翻譯結果可以幫助語意解歧 [Diab et al, 2002][Bhattacharya, 2004]。在本實驗中我們將問題分成兩個部分：a) 如何找出中文詞和英文 詞對應的翻譯，b) 如何解決英文的歧義。 本文接下來的章節組織如下。在第 2 節中說明所使用的資源。第 3 節中說明實驗的方法。第 4 節中說明實驗的結果。結論及未來的發展則在第 5 節中說明。 2. 使用資源 本研究使用了兩本詞典作為語意抽取的對象： a) 國立編譯館學術名詞詞庫 [NICT, 2004]。 b) 英漢詞典 其中國立編譯館所編輯的「學術名詞」詞庫的內容包含 63 個學科類別共 1,046,058 目詞。這些 詞條中有 629,352 目詞是複合詞，佔總詞數的 60 %。英漢詞典共有 208,163 目詞，用來補足 「學術名詞」之不足。此外我們使用 WordNet 2.0 做為語意連結的對象。 由於中文的複合詞在詞和詞之間沒有空白分隔，不像英文詞間以空白字元做為邊界，所以必 需依賴自動斷詞程式將複合詞切分成一般詞。本實驗採用中央研究院詞庫小組所研發的自動斷詞 圖 4. "water tank" 和其組成成份 "tank" 的語意具有上下位關係 . 解歧法二：英文詞之間的語意交集 假設中文詞 w c 可被翻譯成n個不同的英文詞 w e1 , w e2 ,…, w en ，解歧的規則如下： a) w e1 , w e2 ,…, w en 有一個共同的 sy例如，"信號旗" 可翻譯為 "signal", "signal flag" 及 "code flag"，而 "signal" 其中的一個 synset 為 "signal flag" 及 "code flag" 的上位，則 "信號旗" 所標的語意可為 "signal-1", "signal flag-1" 及 "code flag-1"。如圖 5： 2 若 nset s，則w 被連結到 t 為 個 w 位。 signal-1 signal_flag-1 code_flag-1 圖 5. "信號旗" 的英文翻譯 "signal", "signal flag" 及 "code flag" 具有上下位關係 3. 解歧法三：標記沒有歧義的中文詞 利用前面實驗標記所得的結果，找出沒有歧義的中文詞，再利用沒有歧義的中文詞去標 記。例如："防波堤" 在前面實驗所得的結果，都對應到同一個 synset {breakwater-1, groin-2,groyne-1, mole-5, bulwark-3, seawall-1, jetty-1}，所以只要 "防波堤" 對應到的英文詞是該 synset 中的任一詞，就可以判別屬於該 synset。 4. 實驗結果 的對象為「學術名詞詞庫」和「英漢詞典」 ，總共有 1,254,221 個詞，其中所包含 的複合詞有 645,200 佔總詞數的 51.44%。實驗的結果一共將 124,752 個中文詞連結到 42,589 個 WordNet synset，結果一共產生 165,775 個(中文詞, synset) 的連結組合。此結果平均一個中 文詞落在 1.33 (由 165,775 / 124,752 得) 個 WordNet synset 中。在 4.1 節中將說明中英對應的 實驗結果，4.2 節中說明解歧的結果。 .1 中英對應的結果 要對應，所以我們只針對複合詞做評估。評估的方法為在對應好的詞庫 中，隨機抽取 500 個詞條驗證，驗證的方式為人工檢視。結果如表 3 所示，對應的正確率為 95.19 %。 數 對應正確數 正確率 語意抽取 4 由於只有複合詞需 詞條抽樣 500 476 95.19% 表 3. 中英詞與詞對應的正確率 分析這些錯誤的對應約可規類成四種錯誤的類型，如表 4 所示。 。 | 利用雙語學術名詞庫抽取中英字詞互譯及詞義解歧 白明弘 |
d3605195 | We consider the problem of predicting a movie's opening weekend revenue. Previous work on this problem has used metadata about a movie-e.g., its genre, MPAA rating, and cast-with very limited work making use of text about the movie. In this paper, we use the text of film critics' reviews from several sources to predict opening weekend revenue. We describe a new dataset pairing movie reviews with metadata and revenue data, and show that review text can substitute for metadata, and even improve over it, for prediction. | Movie Reviews and Revenues: An Experiment in Text Regression * |
d2032645 | This paper presents our research on automatic question classification for Vietnamese using machine learning approaches. We have experimented with several machine learning algorithms utilizing two kinds of feature groups: bag-of-words and keywords. Our research focuses on two most important tasks which are corpus building and features extraction by crawling data from the Web to build a keyword corpus. The performance of our approach is promising where our system's precision outperforms the state-of-the-art Tree Kernel approach (Collins and Duffy, 2001) on a Vietnamese question corpus. | Learning Based Approaches for Vietnamese Question Classification Using Keywords Extraction from the Web |
d49213156 | With the development of several multilingual datasets used for semantic parsing, recent research efforts have looked into the problem of learning semantic parsers in a multilingual setup(Duong et al., 2017;Susanto and Lu, 2017a). However, how to improve the performance of a monolingual semantic parser for a specific language by leveraging data annotated in different languages remains a research question that is under-explored. In this work, we present a study to show how learning distributed representations of the logical forms from data annotated in different languages can be used for improving the performance of a monolingual semantic parser. We extend two existing monolingual semantic parsers to incorporate such cross-lingual distributed logical representations as features. Experiments show that our proposed approach is able to yield improved semantic parsing results on the standard multilingual GeoQuery dataset. | Learning Cross-lingual Distributed Logical Representations for Semantic Parsing |
d6783611 | The automated scoring of second-language (L2) learner text along various writing dimensions is an increasingly active research area. In this paper, we focus on determining the topical relevance of an essay to the prompt that elicited it. Given the burden involved in manually assigning scores for use in training supervised prompt-relevance models, we develop unsupervised models and show that they correlate well with human judgements.We show that expanding prompts using topically-related words, via pseudo-relevance modelling, is beneficial and outperforms other distributional techniques. Finally, we incorporate our prompt-relevance models into a supervised essay scoring system that predicts a holistic score and show that it improves its performance. | Unsupervised Modeling of Topical Relevance in L2 Learner Text |
d12367328 | This paper presents the adaptation and customization of two lexical resources: Brill tagger, Brill (1992), and EuroWordNet,Vossen et al. (1998), to be used in the ADVICE project devoted to build an intelligent virtual reality sales and service system that uses human language technology. | $GDSWLQJ DQG H[WHQGLQJ OH[LFDO UHVRXUFHV LQ D GLDORJXH V\VWHP |
d250150823 | The aim of this study was to compare the morphological complexity in a corpus representing the language production of younger and older children across different languages. The language samples were taken from the Frog Story subcorpus of the CHILDES corpora, which comprises oral narratives collected by various researchers between 1990 and 2005. We extracted narratives by typically developing, monolingual, middle-class children. Additionally, samples of Lithuanian language, collected according to the same principles, were added. The corpus comprises 249 narratives evenly distributed across eight languages: Croatian, English, French, German, Italian, Lithuanian, Russian and Spanish. Two subcorpora were formed for each language: a younger children corpus and an older children corpus. Four measures of morphological complexity were calculated for each subcorpus: Bane, Kolmogorov, Word entropy and Relative entropy of word structure. The results showed that younger children corpora had lower morphological complexity than older children corpora for all four measures for Spanish and Russian. Reversed results were obtained for English and French, and the results for the remaining four languages showed variation. Relative entropy of word structure proved to be indicative of age differences. Word entropy and relative entropy of word structure show potential to demonstrate typological differences. | Morphological Complexity of Children Narratives in Eight Languages |
d232328187 | International library standards require cataloguers to tediously input Romanization of their catalogue records for the benefit of library users without specific language expertise. In this paper, we present the first reported results on the task of automatic Romanization of undiacritized Arabic bibliographic entries. This complex task requires the modeling of Arabic phonology, morphology, and even semantics. We collected a 2.5M word corpus of parallel Arabic and Romanized bibliographic entries, and benchmarked a number of models that vary in terms of complexity and resource dependence. Our best system reaches 89.3% exact word Romanization on a blind test set. We make our data and code publicly available. | Automatic Romanization of Arabic Bibliographic Records |
d209443709 | While expressions have traditionally been binarized as compositional and noncompositional in linguistic theory, Multiword Expressions (MWEs) demonstrate finer-grained distinctions. Using Association Measures like Pointwise Mutual Information and Dice's Coefficient, MWEs can be characterized as having different degrees of conventionalization and predictability. These gradiences could be reflected during cognitive processing. In this study, fMRI recordings of naturalistic narrative comprehension is used to investigate to what extent these computational measures and their underlying cognitive processes are observable during on-line sentence processing. Our results show that predictability, as quantified through Dice's Coefficent, is a better predictor of neural activation for processing MWEs and the more cognitively plausible computational metric. | Modeling conventionalization and predictability within MWEs at the brain level Anonymous SCiL submission Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE. 200 Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE. 300 Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE. 400 Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE. 500 Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE. 600 Submission ***. Confidential Review Copy. DO NOT DISTRIBUTE |
d3891288 | Determiners, Entities, and Contexts | |
d250390709 | A document can be summarized in a number of ways. Reference-based evaluation of summarization has been criticized for its inflexibility. In this paper, we propose a new automatic reference-free evaluation metric that compares semantic distribution between source document and summary by pretrained language models and considers summary compression ratio. The experiments show that this metric is more consistent with human evaluation in terms of coherence, consistency, relevance, fluency. | Reference-free Summarization Evaluation via Semantic Correlation and Compression Ratio |
d10470801 | For biomedical information extraction, most systems use syntactic patterns on verbs (anchor verbs) and their arguments. Anchor verbs can be selected by focusing on their arguments. We propose to use predicate-argument structures (PASs), which are outputs of a full parser, to obtain verbs and their arguments. In this paper, we evaluated PAS method by comparing it to a method using part of speech (POSs) pattern matching. POS patterns produced larger results with incorrect arguments, and the results will cause adverse effects on a phase selecting appropriate verbs. | Finding Anchor Verbs for Biomedical IE Using Predicate-Argument Structures |
d243948403 | In this paper, we attempt to improve upon the state-of-the-art in predicting a novel's success by modeling the lexical semantic relationships of its contents. We created the largest dataset used in such a project containing lexical data from 17,962 books from Project Gutenberg. We utilized domain specific feature reduction techniques to implement the most accurate models to date for predicting book success, with our best model achieving an average accuracy of 94.0%. By analyzing the model parameters, we extracted the successful semantic relationships from books of 12 different genres. We finally mapped those semantic relations to a set of themes, as defined in Roget's Thesaurus and discovered the themes that successful books of a given genre prioritize. At the end of the paper, we further showed that our model demonstrate similar performance for book success prediction even when Goodreads rating was used instead of download count to measure success. | Syntax and Themes: How Context Free Grammar Rules and Semantic Word Association Influence Book Success |
d9179424 | This paper outlines a formal description of grammatical relations between definitions and verbal predications found in Definitional Contexts in Spanish. It can be situated within the framework of Predication Theory, a model derived from Government & Binding Grammar. We use this model to describe: (i) the syntactic patterns that establish the relationship between definitions and predications; (ii) how useful these patterns are for the identification of definitions in technical corpora. | Workshop On Definition Extraction |
d36651124 | This work deals with the identification of pacients in heterogeneous databases. We deal with the problem of identifier matching as well as with name matching. For that, a method for matching several identification formats was developed; we also developed the QFS measure, which takes in consideration the statistical distribution of names and cultural factors. We show that QFSmeasure is superior to several existing methods for matching names.Resumo. Este trabalho trata da identificação de pacientes em bancos de dados heterogêneos. Lidamos com o problema do emparelhamento de identificadores de pacientes, bem como de nomes. Para tanto, um método de unificação de diversos padrões de identificação e uma medidas de similaridade entre nomes, chamada de medida QFS, foi criada. Mostramos que esta medidaé superior a outros métodos no caso da identificação de nomes de pessoas. * Trabalho desenvolvido e apoiado pelo projeto INCITO CNPq 573589/2008-9. † Bolsista CNPq PQ 302553/2010-0. | Resolução da Heterogeneidade na Identificação de Pacientes * |
d5966798 | In this paper, quantitative analyses of the delay in Japanese-to-English (J-E) and English-to-Japanese (E-J) interpretations are described. The Simultaneous Interpretation Database of Nagoya University (SIDB) was used for the analyses. Beginning time and end time of each word were provided to the corpus using HMM-based phoneme segmentation, and the time lag between the corresponding words was calculated as the word-level delay. Word-level delay was calculated for 3,722 pairs and 4,932 pairs of words for J-E and E-J interpretations, respectively. The analyses revealed that J-E interpretation have much larger delay than E-J interpretation and that the difference of word order between Japanese and English affect the degree of delay. | Construction and Analysis of Word-level Time-aligned Simultaneous Interpretation Corpus |
d5887376 | This paper presents our ongoing work on compilation of English multi-word expression (MWE) lexicon and corpus annotation. We are especially interested in collecting flexible MWEs, in which some other constituents can intervene the expression such as "a number of" vs "a large number of" where a modifier of "number" can be placed in the expression while inheriting the original meaning. We first collect possible candidates of flexible English MWEs from the web, and annotate all of their occurrences in the Wall Street Journal portion of OntoNotes corpus. We make use of word dependency structure information of the sentences converted from the phrase structure annotation. This process enables semi-automatic annotation of MWEs in the corpus and simultaneously produces the internal and external dependency representation of flexible MWEs. | Identification of Flexible Multiword Expressions with the Help of Dependency Structure Annotation |
d564288 | This article outlines a quantitative method for segmenting texts into thematically coherent units. This method relies on a network of lexical collocations to compute the thematic coherence of the different parts of a text from the lexical cohesiveness of their words. We also present the results of an experiment about locating boundaries between a series of concatened texts. | How to thematically segment texts by using lexical cohesion? |
d235258270 | Recent improvements in neural machine translation calls for increased efforts on qualitative evaluations so as to get a better understanding of differences in translation competence between human and machine. This paper reports the results of a study of 1170 adjectives in translation from English to Swedish, using the Parallel Universal Dependencies Treebanks for these languages. The comparison covers two dimensions: the types of solutions employed and the incidence of debatable or incorrect translations. It is found that the machine translation uses all of the solution types that the human translation does, but in different proportions and less competently. | Translation Competence in Machines: A Study of Adjectives in English-Swedish Translation |
d2380594 | We present a coreference resolver called BABAR that uses contextual role knowledge to evaluate possible antecedents for an anaphor. BABAR uses information extraction patterns to identify contextual roles and creates four contextual role knowledge sources using unsupervised learning. These knowledge sources determine whether the contexts surrounding an anaphor and antecedent are compatible. BABAR applies a Dempster-Shafer probabilistic model to make resolutions based on evidence from the contextual role knowledge sources as well as general knowledge sources. Experiments in two domains showed that the contextual role knowledge improved coreference performance, especially on pronouns. | Unsupervised Learning of Contextual Role Knowledge for Coreference Resolution |
d1966007 | Semantic query sub-network is the representation of a natural language query as a graph of semantically connected words. Such sub-networks can be identified as sub-graphs in larger ontologies like DBpedia or Google knowledge graph, which allows for domain and concepts identification, especially in noisy queries. In this paper, we present a novel standalone NLP technique that leverages the cognitive psychology notion of semantic forms for semantic subnetwork extraction from natural language queries. Semantic forms, borrowed from cognitive psychology models, are one of the fundamental structures employed by human cognition to construct semantic information in the brain. We propose a computational cognitive model by means of conditional random fields and explore the interaction patterns among such forms. Our results suggest that the cognitive abstraction provided by semantic forms during labelling can significantly improve parsing and sub-network extraction compared to pure lexical approaches like parts of speech tagging. We conduct experiments on approximately 5000 queries from three diverse datasets to demonstrate the robustness and efficiency of the proposed approach. | A Computational Cognitive Model for Semantic Sub- network Extraction from Natural Language Queries |
d4901838 | The explosion of information technology has led to a substantial growth in quantity, diversity and complexity of linguistic data accessible over the internet. The lack of interoperability between linguistic and language resources represents a major challenge that needs to be addressed, in particular, if information from different sources is to be combined, like, say, machine-readable lexicons, corpus data and terminology repositories. For these types of resources, domain-specific standards have been proposed, yet, issues of interoperability between different types of resources persist, commonly accepted strategies to distribute, access and integrate their information have yet to be established, and technologies and infrastructures to address both aspects are still under development. The goal of the 2nd Workshop on Linked Data in Linguistics (LDL-2013) has been to bring together researchers from various fields of linguistics, natural language processing, and information technology to present and discuss principles, case studies, and best practices for representing, publishing and linking linguistic data collections, including corpora, dictionaries, lexical networks, translation memories, thesauri, etc., infrastructures developed on that basis, their use of existing standards, and the publication and distribution policies that were adopted.Background: Integrating Information from Different SourcesIn recent years, the limited interoperability between linguistic resources has been recognized as a major obstacle for data use and re-use within and across discipline boundaries. After half a century of computational linguistics [8], quantitative typology[12], empirical, corpus-based study of language [10], and computational lexicography[16], researchers in computational linguistics, natural language processing (NLP) or information technology, as well as in Digital Humanities, are confronted with an immense wealth of linguistic resources, that are not only growing in number, but also in their heterogeneity.Interoperability involves two aspects [14]:Structural ('syntactic') interoperability: Resources use comparable formalisms to represent and to access data (formats, protocols, query languages, etc.), i so that they can be accessed in a uniform way and that their information can be integrated with each other.Conceptual ('semantic') interoperability: Resources share a common vocabulary, so that linguistic information from one resource can be resolved against information from another resource, e.g., grammatical descriptions can be linked to a terminology repository. | Linguistic Linked Open Data (LLOD) Introduction and Overview Background: Integrating Information from Differ- ent Sources |
d11806528 | COMPUTATIONAL LINGUISTICS IN 1990 | |
d76655602 | We describe an effort towards the automatic generation of novel riddles in Portuguese, ultimately with humour value. Riddle generation fits in the common architecture of a NLG system and may follow different models, described here, all based on features of a concept, acquired from a lexical-semantic knowledge base. Generated riddles were manually assessed by humans, who rated them as fairly interpretable, surprising, and novel, even if with low humour potential. | Exploring Lexical-Semantic Knowledge in the Generation of Novel Riddles in Portuguese |
d32773716 | Le projet des corpus électroniques de textes en langues mandingues a démarré à St. Petersbourg en 2009. Aujourd'hui, il est effectué par une équipe internationale avec l'implication des spécialistes en langues manding des pays différents. L'outillage tenant compte des caractéristiques spécifiques des langues manding (mais adaptable aux autres langues) a été développé. Le Corpus Bambara de Référence est mis en ligne en 2012, suivi par un corpus maninka (en écriture N'ko et latine) en février 2014. Un correcteur automatique d'orthographe bambara et un logiciel du ROC pour le bambara a été développé sur la base de l'outillage du CBR. L'utilisation expérimentale du CBR dans l'enseignement universitaire du bambara et dans les études linguistiques a montré son efficacité. L'expérience accumulée peut être facilement étendue sur les autres variétés manding (le dioula de RCI, le dioula de Burkina Faso), mais aussi sur d'autres langues africaines. Abstract. The project of electronic corpora for Manding languages was launched in St. Petersburg in 2009. By now, it is carried out by an international team with an assistance by specialists in Manding languages from different countries. Tools have been developed taking into account the specifics of Manding languages (and adaptable to other languages). The Bamana Reference Corpus was put on line in 2012, it was followed by a Maninka corpus (in both Roman and N'ko writing) in February 2014. An orthography corrector for Bamana and a software for the Bamana OCR has been developed on the basis of the Bamana Reference Corpus tools. An experimental use of the Bamana Corpus in the Bamana teaching in universities and in linguistic studies has proved its effectiveness. The experience accumulated in the framework of this project can be relatively easily extended to other Manding varieties (Jula of Côte d'Ivoire, Jula of Burkina Faso), and, if necessary, to other African languages. | Projet des corpus écrits des langues manding : le bambara, le maninka 1 |
d16129280 | Aims of the projectThe general aim of our project is to improve the quality of existing systems extracting knowledge from texts by introducing refined lexical semantics data. The conlribution of lexical ~mantics to knowledge extraction is not new and has already been demonstrated in a few systems. Our more precise aims are to:propose and show feasability of more radical semantic classifications which facilitate lexical descriptions by factoring out as much information as possible, enhancing re-usability of linguistic ressources. We show how the different linguistic ressources can be org~mized and how they interact, -investigate different levels of granularity in the semantic descriptions and their impact on the quality of the extracted knowledge. In our system, granularity is considered at two levels: (1) linguistic: linguistic knowledge representations may be more or less precise, (2) functional: most modules of our system can work independently and thus can be used ~pamtely, -evaluate different algorithms for extracting knowledge, taking into account efficiency aspects, -evaluate the costs of extending our system to larger sets of texts anti to differeut application domains. Our prqiect is applied to research projects descriptions (noted hereafter as RPD) where the annual work of researchers at the DER of EDF (Direction des Etudes et des Rechcrches, Electricit6 de France) is described in terms of research actions. The extracted knowledge must be sufficiently accurate to allow for the realization of the following Imrposes: (1) evaluation of the importance of the use of techniques, procedures anti equipments, (2) automatic distribution of documents in different services, (3) interrogation, e.g. who does what anti what kind of results are available, (4) identification of relations of various types between projects, (5) construction of synthesis of research activities on precise topics, and (6) creation of the 'history' of a project.About 2.000 RPD are produced each year, each of about 200 words hmg. The total vocabulary is about 50.000 different words. Texts include fairly complex linguistic constructs. We also use the EDF thesaurus (encoding for nouns: taxonomies, associative relations, and synonyms, in a broad sense).In this document, we first introduce the linguistic organization of our project, present the general form of texts and identify the type of information which mnst be extracted out of them. Next, we present a semantic representation for the extracted knowlexlge, and study in more depth the extraction of information under the form of predicate-argument and predicate-modifier structures (Jackendoff 87a, Ka~ and Fodor 63). | Knowledge Extraction from Texts: a method for extracting predicate-argument structures from texts |
d199577634 | To advance models of multimodal context, we introduce a simple yet powerful neural architecture for data that combines vision and natural language. The "Bounding Boxes in Text Transformer" (B2T2) also leverages referential information binding words to portions of the image in a single unified architecture. B2T2 is highly effective on the Visual Commonsense Reasoning benchmark 1 , achieving a new state-of-the-art with a 25% relative reduction in error rate compared to published baselines and obtaining the best performance to date on the public leaderboard (as of May 22, 2019). A detailed ablation analysis shows that the early integration of the visual features into the text analysis is key to the effectiveness of the new architecture. A reference implementation of our models is provided 2 . | Fusion of Detected Objects in Text for Visual Question Answering |
d5887608 | Internet communication plays a considerable part in economic, financial and even politic domains. It is greatly influencing the politic revolution of many Arabic countries. That allows Internet communication to take more and more scale especially in an Arabic context. In this case, we notice that Internet communication is based on textual interchange using Arabic dialects more than Arabic language. However, few efforts were made for Arabic dialect processing particularly for aeb 1 language. In this case, we suggest building a standardized aeb Wordnet, which is a basic tool for Natural Language Processing (NLP) of aeb language. In this article, we present an extended Wordnet-LMF model acquired to aeb language specificities used to represent aeb Wordnet and we describe building steps. | Building a standardized Wordnet in the ISO LMF for aeb language Hsan Soussou(2) |
d33924822 | Recently, language resources (LRs) are becoming indispensable for linguistic researches. However, existing LRs are often not fully utilized because their variety of usage is not well known, indicating that their intrinsic value is not recognized very well either. Regarding this issue, lists of usage information might improve LR searches and lead to their efficient use. In this research, therefore, we collect a list of usage information for each LR from academic articles to promote the efficient utilization of LRs. This paper proposes to construct a text corpus annotated with usage information (UI corpus). In particular, we automatically extract sentences containing LR names from academic articles. Then, the extracted sentences are annotated with usage information by two annotators in a cascaded manner. We show that the UI corpus contributes to efficient LR searches by combining the UI corpus with a metadata database of LRs and comparing the number of LRs retrieved with and without the UI corpus. | Collection of Usage Information for Language Resources from Academic Articles |
d248780029 | Goal-oriented dialogues generation grounded in multiple documents(MultiDoc2Dial) is a challenging and realistic task. Unlike previous works which treat document-grounded dialogue modeling as a machine reading comprehension task from single document, Multi-Doc2Dial task faces challenges of both seeking information from multiple documents and generating conversation response simultaneously. This paper summarizes our entries to agent response generation subtask in Multi-Doc2Dial dataset. We propose a three-stage solution, Grounding-guided goal-oriented dialogues generation(G4), which predicts groundings from retrieved passages to guide the generation of the final response. Our experiments show that G4 achieves SacreBLEU score of 31.24 and F1 score of 44.6 which is 60.7% higher than the baseline model. | G4: Grounding-guided Goal-oriented Dialogues Generation with Multiple Documents |
d52125956 | In this paper, we describe Mumpitz, the system we submitted to the PARSEME Shared Task on automatic identification of verbal multiword expressions (VMWEs). Mumpitz consists of a Bidirectional Recurrent Neural Network (BRNN) with Long Short-Term Memory (LSTM) units and a heuristic that leverages the dependency information provided in the PARSEME corpus data to differentiate VMWEs in a sentence. We submitted results for seven languages in the closed track of the task and for one language in the open track. For the open track we used the same system, but with pretrained instead of randomly initialized word embeddings to improve the system performance.This work is licensed under a Creative Commons Attribution 4.0 International License. License details: http: //creativecommons.org/licenses/by/4.0/ | Mumpitz at PARSEME Shared Task 2018: A Bidirectional LSTM for the Identification of Verbal Multiword Expressions |
d8337297 | The relative frequencies of character bigrams appear to contain much information for predicting the first language (L1) of the writer of a text in another language (L2). Tsur and Rappoport(2007)interpret this fact as evidence that word choice is dictated by the phonology of L1. In order to test their hypothesis, we design an algorithm to identify the most discriminative words and the corresponding character bigrams, and perform two experiments to quantify their impact on the L1 identification task. The results strongly suggest an alternative explanation of the effectiveness of character bigrams in identifying the native language of a writer. | Does the Phonology of L1 Show Up in L2 Texts? |
d5116525 | This paper describes measures for evaluating the three determinants of how well a probabilistic elassifter performs on a given test set. These determinants are the appropriateness, for the test set, of the results of (1)feature selection, (2) formulation of the parametric form of the model, and(3)parameter estimation. These are part of any model formulation procedure, even if not broken out as separate steps, so the tradeoffs explored in this paper are relevant to a wide variety of methods. The measures are demonstrated in a large experiment, in which they are used to analyze the results of roughly 300 classifiers that perform word-sense disambiguation. | The Measure of a Model * |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.