Syntactic Similarity Research Articles

Syntactic and pragmatic transfer effects in reported-speech constructions in three contact varieties of English influenced by Afrikaans

Cross-linguistic influence (CLI) is investigated by various research traditions that examine language contact varieties. This article presents an analysis of syntactic and pragmatic effects of CLI in three corpora of published, written varieties of English influenced by contact with Afrikaans: native White South African English, English second-language writing by Afrikaans speakers, and texts translated from Afrikaans into English. These varieties differ in the strength of bilingual activation as well as in different sociocognitive conditions of production. Three related reported-speech constructions that exhibit syntactic and pragmatic similarities and differences are analysed, to determine how bilingual activation and sociocognitive factors interact with CLI. The findings indicate that overt syntactic transfer is almost completely absent in the language of highly proficient bilinguals, but covert syntactic transfer takes place where an Afrikaans construction corresponds to only one of two variants available to native varieties of English. Extensive evidence of covert pragmatic transfer is found across various registers. The cognitive strain associated with high levels of bilingual activation overrides syntactic and pragmatic CLI for one construction investigated, in favour of a consistent preference for the explicit and more formal variant. This effect is further enhanced by the selection of the more formal variant during the process of post-production editing.

How are functionally similar code clones syntactically different? An empirical study and a benchmark

Background. Today, redundancy in source code, so-called “clones” caused by copy&paste can be found reliably using clone detection tools. Redundancy can arise also independently, however, not caused by copy&paste. At present, it is not clear how onlyfunctionally similar clones(FSC) differ from clones created by copy&paste. Our aim is to understand and categorise the syntactical differences in FSCs that distinguish them from copy&paste clones in a way that helps clone detection research.Methods. We conducted an experiment using known functionally similar programs in Java and C from coding contests. We analysed syntactic similarity with traditional detection tools and explored whether concolic clone detection can go beyond syntax. We ran all tools on 2,800 programs and manually categorised the differences in a random sample of 70 program pairs.Results. We found no FSCs where complete files were syntactically similar. We could detect a syntactic similarity in a part of the files in <16% of the program pairs. Concolic detection found 1 of the FSCs. The differences between program pairs were in the categories algorithm, data structure, OO design, I/O and libraries. We selected 58 pairs for an openly accessible benchmark representing these categories.Discussion. The majority of differences between functionally similar clones are beyond the capabilities of current clone detection approaches. Yet, our benchmark can help to drive further clone detection research.

A Learning-Based Fact Selector for Isabelle/HOL

Sledgehammer integrates automatic theorem provers in the proof assistant Isabelle/HOL. A key component, the fact selector, heuristically ranks the thousands of facts (lemmas, definitions, or axioms) available and selects a subset, based on syntactic similarity to the current proof goal. We introduce MaSh, an alternative that learns from successful proofs. New challenges arose from our zero click vision: MaSh integrates seamlessly with the users' workflow, so that they benefit from machine learning without having to install software, set up servers, or guide the learning. MaSh outperforms the old fact selector on large formalizations.

Exploiting Syntactic Similarities for Preposition Error Corrections on Indonesian Sentences Written by Second Language Learner

We propose a method to artificially generate training data to correct preposition errors in Indonesian sentences written by second language learners. Basically, we injected large size of native sentences with preposition errors learned from learners’ sentences. Our method copies a preposition error from a learner sentence to a native sentence by firstly calculating a syntactic similarity score between the native sentence and the learners’ sentence. Then, it chooses the preposition error from the learner sentence that has the highest syntactic similarity score to the native sentence, to replace the original preposition in the native sentence.Experimental results show that the preposition error correction model trained on the artificial data resulted from our method outperforms the correction model trained on the similar size of native data.

Boosting the Quality of Approximate String Matching by Synonyms

A string-similarity measure quantifies the similarity between two text strings for approximate string matching or comparison. For example, the strings “Sam” and “Samuel” can be considered to be similar. Most existing work that computes the similarity of two strings only considers syntactic similarities, for example, number of common words or q -grams. While this is indeed an indicator of similarity, there are many important cases where syntactically-different strings can represent the same real-world object. For example, “Bill” is a short form of “William,” and “Database Management Systems” can be abbreviated as “DBMS.” Given a collection of predefined synonyms, the purpose of this article is to explore such existing knowledge to effectively evaluate the similarity between two strings and efficiently perform similarity searches and joins, thereby boosting the quality of approximate string matching. In particular, we first present an expansion-based framework to measure string similarities efficiently while considering synonyms. We then study efficient algorithms for similarity searches and joins by proposing two novel indexes, called SI-trees and QP-trees, which combine signature-filtering and length-filtering strategies. In order to improve the efficiency of our algorithms, we develop an estimator to estimate the size of candidates to enable an online selection of signature filters. This estimator provides strong low-error, high-confidence guarantees while requiring only logarithmic space and time costs, thus making our method attractive both in theory and in practice. Finally, the experimental results from a comprehensive study of the algorithms with three real datasets verify the effectiveness and efficiency of our approaches.

The improvement of question process method in Q&A system

Life service information plays an important role in people’s life, such as weather conditions, so the study of how to get lifeservice information has important significance. This paper put forward a question processing method called “integrated semanticalgorithm” in Q&A System of life service information. The new algorithm was based on the semantic web, word order similarityalgorithm and the syntactic similarity algorithm. When matching the question templates, especially for some question templateswhich are characteristic of certain fields, the new algorithm can identify the type of questions, narrow the matching range of thequestion templates, and improve the matching accuracy. In the experiment, we chose “weather field” as the experimental subject.In the first experiment, we built the question syntactic templates and semantic web of weather, and collected 55 questions ofweather title as test set. Then we used the word similarity algorithm, the syntactic similarity algorithm and integrated semanticsimilarity algorithm to match question templates with the test question set. The experimental results show that the integratedsemantic algorithm is better than the other two algorithms in matching accuracy. In the second experiment, we randomlyselected some questions from different fields, then we used the three similarity algorithms in the first experiment to do the fielddistinguishing experiment. The experiment shows that only the integrated semantic algorithm can recognize questions of differentfields.

Query-based multi-documents summarization using linguistic knowledge and content word expansion

In this paper, a query-based summarization method, which uses a combination of semantic relations between words and their syntactic composition, to extract meaningful sentences from document sets is introduced. The problem with current statistical methods is that they fail to capture the meaning when comparing a sentence and a user query; hence there is often a conflict between the extracted sentences and users' requirements. However, this particular method can improve the quality of document summaries because it is able to avoid extracting a sentence whose similarity with the query is high but whose meaning is different. The method is executed by computing the semantic and syntactic similarity of the sentence-to-sentence and sentence-to-query. To reduce redundancy in summary, this method uses the greedy algorithm to impose diversity penalty on the sentences. In addition, the proposed method expands the words in both the query and the sentences to tackle the problem of information limit. It bridges the lexical gaps for semantically similar contexts that are expressed using different wording. The experimental results display that the proposed method is able to improve performance compared with the participating systems in DUC 2006. The experimental results also showed that the proposed method demonstrates better performance as compared to other existing techniques on DUC 2005 and DUC 2006 datasets.

A Hybrid Layered Approach for Ontology Matching

Ontologies act as a key technology for the visualization of the semantic web and their existence in different domains produces a heterogeneity problem for information integration. The process of ontology matching solves the heterogeneity problem by identifying the semantically related identical entities available from different ontologies. Finding similarity between the class names across the input ontologies is the major step performed in every ontology matching process. Most of the existing methods compare the class names only based on their similarity irrespective of their real meaning. This paper suggests a hybrid layered approach for matching the ontology classes. This approach calculates syntactic, semantic and structural similarity between the classes from input ontologies in successive layers. Finally, the alignment layer generates the final matching results by combining the results obtained from the previous layers and generates semantic mappings between them. This approach can also be applied to match other entities of the ontology. The results obtained prove that the proposed approach overtakes the other existing methods, thereby improving the accuracy of the results achieved.

Learning English in Algeria Through French-based Background Proficiency

With globalization leading to an ever-increasing connectedness of the world, learning a foreign language has become a requirement rather than an indulgence. Yet, learning a second and/or a third language may present a challenge and the issue in Algeria is no exception. In addition to the mother tongue (Arabic or Berber in some regions), French is the second most spoken language in the country, and thus, its relatedness with other Western languages facilitates the learning of English as a third language. Already acquired, not only in the primary school process but also in the community's linguistic practices, French helps learners to develop proficiency in English through the transfer of their French-based background, in particular the typological (syntactical and lexical) similarities between French and English. The aim of the study is to show that the learners’ prior linguistic experience in L2 may be a facilitator to learn an L3. In this study, a group of 25 male learners aged between 20 and 30 years old was chosen. Their L1 is Arabic. The participants are bilinguals (Arabic/French) with different levels of competence in French. They had to translate an English text into the language of their choice, i.e., either Arabic or French. The results have shown that learners who already know French find it easy to learn English because of their ability to use resources in L2 to learn L3.

PDLK: Plagiarism detection using linguistic knowledge

Plagiarism is described as the reuse of someone else's previous ideas, work or even words without sufficient attribution to the source. This paper presents a method to detect external plagiarism using the integration of semantic relations between words and their syntactic composition. The problem with the available methods is that they fail to capture the meaning in comparison between a source document sentence and a suspicious document sentence, when two sentences have same surface text (the words are the same) or they are a paraphrase of each other. Therefore it causes inaccurate or unnecessary matching results. However, this method can improve the performance of plagiarism detection because it is able to avoid selecting the source text sentence whose similarity with suspicious text sentence is high but its meaning is different. It is executed by computing the semantic and syntactic similarity of the sentence-to-sentence. Besides, the proposed method expands the words in sentences to tackle the problem of information limit. It bridges the lexical gaps for semantically similar contexts that are expressed in a different wording. This method is also capable to identify various kinds of plagiarism such as the exact copied text, paraphrasing, transformation of sentences and changing of word structure in the sentences. As a result, the experimental results have displayed that the proposed method is able to improve the performance compared with the participating systems in PAN-PC-11. The experimental results also displayed that the proposed method demonstrates better performance as compared to other existing techniques on PAN-PC-10 and PAN-PC-11 datasets.

Semantic Similarity from Natural Language and Ontology Analysis

Artificial Intelligence federates numerous scientific fields in the aim of developing machines able to assist human operators performing complex treatments -- most of which demand high cognitive skills (e.g. learning or decision processes). Central to this quest is to give machines the ability to estimate the likeness or similarity between things in the way human beings estimate the similarity between stimuli. In this context, this book focuses on semantic measures: approaches designed for comparing semantic entities such as units of language, e.g. words, sentences, or concepts and instances defined into knowledge bases. The aim of these measures is to assess the similarity or relatedness of such semantic entities by taking into account their semantics, i.e. their meaning -- intuitively, the words tea and coffee, which both refer to stimulating beverage, will be estimated to be more semantically similar than the words toffee (confection) and coffee, despite that the last pair has a higher syntactic similarity. The two state-of-the-art approaches for estimating and quantifying semantic similarities/relatedness of semantic entities are presented in detail: the first one relies on corpora analysis and is based on Natural Language Processing techniques and semantic models while the second is based on more or less formal, computer-readable and workable forms of knowledge such as semantic networks, thesaurus or ontologies. (...) Beyond a simple inventory and categorization of existing measures, the aim of this monograph is to convey novices as well as researchers of these domains towards a better understanding of semantic similarity estimation and more generally semantic measures.

A multimodal feature learning approach for sentiment analysis of social network multimedia

In this paper we investigate the use of a multimodal feature learning approach, using neural network based models such as Skip-gram and Denoising Autoencoders, to address sentiment analysis of micro-blogging content, such as Twitter short messages, that are composed by a short text and, possibly, an image. The approach used in this work is motivated by the recent advances in: i) training language models based on neural networks that have proved to be extremely efficient when dealing with web-scale text corpora, and have shown very good performances when dealing with syntactic and semantic word similarities; ii) unsupervised learning, with neural networks, of robust visual features, that are recoverable from partial observations that may be due to occlusions or noisy and heavily modified images. We propose a novel architecture that incorporates these neural networks, testing it on several standard Twitter datasets, and showing that the approach is efficient and obtains good classification results.

Architecture of DSSim

This paper presents the architecture of DSSim (DSSim stands for Similarity based on Dempster-Shafer) our multi-agent ontology mapping system. It describes several types of agents and their roles in the DSSim architecture. These agents are mapping agents which are able to perform either semantic or syntactic similarity. The authors' architecture is generic as no mappings need to be learned in advance and it could be easily extended by adding new mapping agents in the framework. These new mapping agents could run different similarity algorithms either semantic or syntactic. In this way, DSSim could assess which algorithm has a better performance. Additionally, this paper presents the main algorithms used in DSSim and discusses DSSim advantages and drawbacks.

Affirmation without End: Some Syntactical Similarities between the Poetry of Yves Bonnefoy and the Philosophy of Jean-Luc Nancy

Affirmation without End: Some Syntactical Similarities between the Poetry of Yves Bonnefoy and the Philosophy of Jean-Luc Nancy Emily McLaughlin SINCE HE CAME TO PROMINENCE with the publication of Du mouvement et de l’immobilité de Douve in 1953, Yves Bonnefoy has always been quick to affirm the value and the function of poetry. Whether writing about his own work or the work of others, whether writing essays, poetry or prose, he reminds us that poetry’s function is affirmative. Poetry is presented as an ontological act, an affirmation of the sense of existence, and a quest to affirm the physical and embodied reality of “la présence” within language. In “L’acte et le lieu de la poésie,” published in L’improbable in 1959, the poet insists that the “destin privé de chaque œuvre” is to “donner un sens à ce qui est.”1 The poet continues, “je ne crois pas que soit de poésie vraie que ne cherche aujourd’hui, et ne veuille chercher jusqu’au dernier souffle, à fonder un nouvel espoir” (Bonnefoy, L’improbable 122–23). It is therefore no surprise that Bonnefoy’s writings are dotted with simple affirmative statements, statements that use the verb “être” to affirm the existence of a phenomenon and then provide a simple definition of it. Bonnefoy writes in the early seminal essay “Les tombeaux de Ravenne, “L’objet sensible est présence” (Bonnefoy, L’improbable 26); “Simplement, [le monde] est avec nous. Dans le sensible” (Bonnefoy, L’improbable 28). In an essay on Paul Valéry, he affirms, “La poésie comme l’amour doit décider que les êtres sont” (Bonnefoy, L’improbable, 101). In a more recent essay on Christian Dotremont, published in 2010, he affirms, “la vie est une réalité ultime, enracinée si profond dans l’être qu’elle s’y obstine en dépit de tous les obstacles, qu’elle est donc comme telle la vérité, mais une vérité âpre, ‘rugueuse.’”2 As his use of the word “rugueuse” here suggests, Bonnefoy conceives of this act of affirmation in Rimbaldien terms. The poet is the one who affirms “la réalité rugueuse à étreindre,” in all its irreducibility and ungraspability.3 And he does so, over and over again. Critics have been quick to discern the provocative or resistant nature of Bonnefoy’s affirmations. Roland Gérard Giguère and Michèle Finck have highlighted how Bonnefoy’s affirmative poetics advertise his fidelity to material reality or embodied experience,4 at a time when writers and critics are increasingly [End Page 110] preoccupied with the modalities of linguistic production, the ways in which language forms our perception and experience of reality. With the emergence from the 1950s onwards of structuralism and post-structuralism, and literary and psychoanalytic strains of theory, Bonnefoy’s simple affirmations seek to counter what he describes as “un parti pris des mots” (Bonnefoy, Le siècle 174). The poet affirms the primacy of presence, the real, and the world. He affirms the sense of words that Heidegger and then Derrida place scrupulously under erasure, words that according to a logic of différance are irretrievably spaced and suspended. At a time of growing artistic and critical skepticism, Bonnefoy’s affirmations appear to be defiantly or even willfully naive. Yet Bonnefoy’s affirmations are so striking precisely because they are so effusive. An initial affirmation —“Le sensible est une présence,” for example—offers a definition that is simple and persuasive, assuaging a basic human need for clarity and identity. This definition then tends to be supplemented by an appositional syntax that repeatedly equates one term with another. The more enthusiastic the poet’s affirmations grow, the more they tend to spill, from one clause to another, in a long series of appositions, parataxes or parentheses. The poet’s effusively affirmative syntax thus voids processes of definition. Bonnefoy affirms the existence of phenomena that cannot be defined, measured or signified. And whilst the words that he affirms are not erased, as they are in Heidegger’s or Derrida’s writings, they are nonetheless revealed to be inherently elusive or excessive. Bonnefoy writes, “Le sensible est une présence...

Computation of Program Source Code Similarity by Composition of Parse Tree and Call Graph

This paper proposes a novel method to compute how similar two program source codes are. Since a program source code is represented as a structural form, the proposed method adopts convolution kernel functions as a similarity measure. Actually, a program source code has two kinds of structural information. One is syntactic information and the other is the dependencies of function calls lying on the program. Since the syntactic information of a program is expressed as its parse tree, the syntactic similarity between two programs is computed by a parse tree kernel. The function calls within a program provide a global structure of a program and can be represented as a graph. Therefore, the similarity of function calls is computed with a graph kernel. Then, both structural similarities are reflected simultaneously into comparing program source codes by composing the parse tree and the graph kernels based on a cyclomatic complexity. According to the experimental results on a real data set for program plagiarism detection, the proposed method is proved to be effective in capturing the similarity between programs. The experiments show that the plagiarized pairs of programs are found correctly and thoroughly by the proposed method.

AndroSimilar: Robust signature for detecting variants of Android malware

Android Smartphone popularity has increased malware threats forcing security researchers and AntiVirus (AV) industry to carve out smart methods to defend Smartphone against malicious apps. Robust signature based solutions to mitigate threats become necessary to protect the Smartphone and confidential user data. Here we present AndroSimilar, an approach which generates signatures by extracting statistically robust features, to detect malicious Android apps. Proposed method is effective against code obfuscation and repackaging, widely used techniques to propagate unseen variants of known malware by evading AV signatures. AndroSimilar is a syntactic foot-printing mechanism that finds regions of statistical similarity with known malware to detect those unknown, zero day samples. We also show that syntactic similarity considering whole app, rather than just embedded DEX file is more effective, contrary to known fuzzy hashing approach. We also apply clustering algorithm to identify small set of family signatures to reduce overall signature database size. Proposed approach can be refined to deploy as Smartphone AV.

Effects of representational distance between meanings on the neural correlates of semantic ambiguity

Psycholinguistic research demonstrates that representational distance between meanings influences recognition of ambiguous words. Our goal was to investigate whether the neural correlates of ambiguity are also modulated by representational distance as a function of syntactic similarity (i.e., grammatical class) and meaning dominance. In an event-related fMRI experiment, participants completed a visual lexical decision task that included balanced and unbalanced noun–noun and noun–verb homonyms, unambiguous words, and nonwords. Syntactic similarity effects were observed in left inferior frontal regions, with greater activation for noun–verb than noun–noun homonyms. Meaning dominance effects were observed in left middle and superior temporal regions, with greater activation for balanced than unbalanced homonyms. These findings indicate that the behavioral cost associated with processing ambiguous word meanings, modulated by syntactic similarity and meaning dominance, is reflected in the neural systems underlying ambiguity processing, as frontal and temporal regions are recruited by increased competition as a function of representational distance.

Influence of Second Language Proficiency and Syntactic Structure Similarities on the Sensitivity and Processing of English Passive Sentence in Late Chinese-English Bilinguists: An ERP Study.

To investigate the influence of L2 proficiency and syntactic similarity on English passive sentence processing, the present ERP study asked 40 late Chinese-English bilinguals (27 females and 13 males, mean age=23.88) with high or intermediate L2 proficiency to read the sentences carefully and to indicate for each sentence whether or not it was correct. Sentences were classified into literal translation sentences with the similar structure between the two languages and free translation sentences with the different structure. Behavioral data showed: shorter reaction times and higher accuracy rates occurred in the high-proficient group than those in the intermediate-proficient group; shorter reaction times and higher accuracy rates were observed in literal translation sentences than those in free translation sentences. ERP results showed literal translation sentences elicited an enhanced P200 and P600 while free translation sentences elicited a larger N400. The high-proficient group showed a larger P600 in syntactic violations and double violations while the intermediate-proficient group evoked an enhanced N400 in semantic violations and double violations. Literal translation sentences caused a larger P200 while free translation sentences elicited more negative-going N400. Behavioral and ERP data revealed the influence of L2 proficiency and syntactic similarity on L2 sentence processing, and L2 proficiency played a predominate role.

A unified analysis of conditionals as topics

We bring out syntactic and semantic similarities of two types of conditionals with fronted antecedents [normal indicative conditionals (NCs) and biscuit conditionals (BCs)] and two types of left dislocation constructions in German (German left dislocation and hanging topic left dislocation), which mark two types of topicality (aboutness topicality and relevance topicality). On the basis of these similarities we argue that (the antecedent if-clauses of) NCs and BCs are aboutness topics and relevance topics, respectively. Our analysis extends the approach to aboutness topicality of Endriss (Quantificational topics. A scopal treatment of exceptional wide scope phenomena, 2009) to relevance topics to derive the semantic and pragmatic contribution of left-dislocated DPs and applies it to an analysis of conditionals as (maximal) pluralities of possible worlds. We show how this uniform approach to the interpretation of topicality accounts for the nominal left dislocation constructions as well as for the semantic and pragmatic effects observed in connection with the two types of conditionals. We furthermore discuss the potential of our proposal to deal with subjunctive biscuit conditionals, if-clauses modifying speech acts different from assertions, conditionals with right-dislocated if-clauses, and nested conditionals.

Rewarding Research

Rewarding Research