Search

Each programming language needs a mechanism of translating source code into code understood by the computer, i.e., machine code or assembly. This is done by a compiler. A compiler is thus a fundamental tool for software development. Lexical analysis, as the first step of compilation, has an important role and should be performed in an optimal way. But not only compilers rely on lexical analysis. Its role lies nowadays also in pattern recognition or natural language processing. In this paper, we present some of the main aspects of lexical analysis and how to efficiently perform it. Our aim is to offer some guidelines in constructing a modern and efficient scanner (lexer), outlining the criteria, which should be considered, presenting some pitfalls, and comparing existent tools for compiler construction. The main contribution of the paper is to offer practical guidelines and recommendations, especially to graduate students and programmers, in constructing an efficient scanner (lexer), considering also the proper tools currently available. [ABSTRACT FROM AUTHOR]

Requirement traceability links are an essential part of requirement management software and are a basic prerequisite for software artifact changes. The manual establishment of requirement traceability links is time-consuming. When faced with large projects, requirement managers spend a lot of time in establishing relationships from numerous requirements and codes. However, existing techniques for automatic requirement traceability link recovery are limited by the semantic disparity between natural language and programming language, resulting in many methods being less accurate. In this paper, we propose a fine-grained requirement-code traceability link recovery approach based on query expansion, which analyzes the semantic similarity between requirements and codes from a fine-grained perspective, and uses a query expansion technique to establish valid links that deviate from the query, so as to further improve the accuracy of traceability link recovery. Experiments showed that the approach proposed in this paper outperforms state-of-the-art unsupervised traceability link recovery methods, not only specifying the obvious advantages of fine-grained structure analysis for word embedding-based traceability link recovery, but also improving the accuracy of establishing requirement traceability links. The experimental results demonstrate the superiority of our approach. [ABSTRACT FROM AUTHOR]

In this world of modern data, it is so difficult to recognize handwritten characters for Tamil as many people have different styles of writing, so some of the letters are very difficult to understand and only a few can understand them. So, to overcome this issue, we built an algorithm in which the system could recognize the character and return the output. As it is difficult to understand letters manually for all their text, there is a need for some automatic method. The only intention of character recognition is that it wants to create a high-quality, accurate result that has the important points while considering the outlined input source image. Mostly, natural language processing and machine learning face the same problem with text recognition. The main goal of automatic character recognition is to create a high degree of accuracy as best as a human can do. Character recognition is the process of filtering the required information from the input-trained source to output the most useful content. This paper proposes a CNN-VGG16-RF model (convolution neural network-VGGNet-random forest) which employs an effective method to pick out the correct output. Experimental tests for our model were carried out to evaluate text quality, and the Tamil language dataset from the HP Tamil Lab website was used to compare our model to some other models; our model was found to be more effective in solving the handwritten recognition problem. In this model, we are going to propose Tamil vowels such as 12 letters only for the training and testing process. [ABSTRACT FROM AUTHOR]

In recent years, multimodal learning has gained acceptability because of the availability of low resource-consuming fusion techniques and robust and powerful deep learning architectures. Visual question answering (VQA) is an interdisciplinary research domain in natural language processing and computer vision. In previous works, researchers have tried to optimize the VQA problems primarily with the help of optimized bilinear fusion techniques. In this paper, we propose a novel question segregation framework for visual question answering to optimize the VQA problem where the VQA framework is segregated by the question type labels. The main contribution of the proposed question segregation framework is the reduction in the execution time and computational resource requirement of VQA models. Six VQA models are tested under the proposed framework and have promising results. The proposed question segregation framework can be extended to other VQA models and datasets, and the problem of model bias toward larger volume question type labels is rectified by having an individual model for each question type label. Also, the proposed question segregation framework provides meaningful answers for a given question by filtering out unrealistic answers by restricting the answer space. [ABSTRACT FROM AUTHOR]

The leading intention of the current paper is to review the research work accomplished by various researchers to achieve sentiment analysis on the text and to elaborate on natural language processing (NLP) and various machine learning algorithms used to evaluate textual sentiments. In this study, primitive cases are considered that used crucial algorithms, and knowledge that can be opted for sentiment analysis. A survey of the work that has been done till now is conducted observing the results and outcomes concerning varying parameters of various researchers who worked on previously existing as well as novel and hybrid algorithms opting legion methodologies. The fundamental algorithms like Support Vector Machine (SVM), Bayesian Networks (BN), Maximum Entropy (MaxEnt), Conditional Random Fields (CRF) and Artificial Neural Networks (ANN) are also discussed to achieve practice percentage and accuracy score in the field of NLP, sentiment analysis and text analytics. Various other novel approaches and algorithms like CNN, LSTM, KNN, K*, K‐means, K‐means++, SOM and ENORA, along with their limitations and the performance metrics providing accuracies for major open data sets are also analyzed. [ABSTRACT FROM AUTHOR]

Research of visual neural networks (VNNs) is one of the most important topics in deep learning and has received wide attention from industry and academia for their promising performance. The applications of VNNs range from image classification and target detection to scene segmentation in various fields such as transportation, healthcare and finance. In general, VNNs can be divided into two types: Convolutional neural networks (CNNs) and Transformer networks. In the last decade, CNNs have dominated the research of vision tasks. Recently, Transformer networks are successfully used in the fields of natural language processing and computer vision, and have achieved remarkable performance in many vision tasks. In this paper, the basic architectures and current trends of these two types of VNNs are first introduced. Then, three major challenges of VNNs are pointed out: scalability, robustness and interpretability. Next, the lightweight, robust and interpretable solutions are summarized and analyzed. Finally, the future opportunities of VNNs are presented. [ABSTRACT FROM AUTHOR]

Despite the prevalence of disagreement between users on social media platforms, studies of online debates typically only look at positive online interactions, represented as networks with positive ties. In this paper, we hypothesize that the systematic neglect of conflict that these network analyses induce leads to misleading results on polarized debates. We introduce an approach to bring in negative user-to-user interaction, by analyzing online debates using signed networks with positive and negative ties. We apply this approach to the Dutch Twitter debate on 'Black Pete'—an annual Dutch celebration with racist characteristics. Using a dataset of 430,000 tweets, we apply natural language processing and machine learning to identify: (i) users' stance in the debate; and (ii) whether the interaction between users is positive (supportive) or negative (antagonistic). Comparing the resulting signed network with its unsigned counterpart, the retweet network, we find that traditional unsigned approaches distort debates by conflating conflict with indifference, and that the inclusion of negative ties changes and enriches our understanding of coalitions and division within the debate. Our analysis reveals that some groups are attacking each other, while others rather seem to be located in fragmented Twitter spaces. Our approach identifies new network positions of individuals that correspond to roles in the debate, such as leaders and scapegoats. These findings show that representing the polarity of user interactions as signs of ties in networks substantively changes the conclusions drawn from polarized social media activity, which has important implications for various fields studying online debates using network analysis. [ABSTRACT FROM AUTHOR]

Natural language processing (NLP) is an area in artificial intelligence that deals with understanding, interpretation and development of human language for computers to carry out tasks such as sentiment analysis, summarization of text in a document, developing conversational agents, machine translation and speech recognition. From conversational agents called catboats deployed on various websites that interact with consumers digitally to understand the needs of the consumers to reading summarized content delivered through apps in smartphones, NLP has had some major achievements in transforming the digital world that is increasingly gearing towards artificial intelligence. One area that has seen remarkable growth in recent times is language modelling, a statistical technique to compute the probability of tokens or words in a given sentence. In this paper, we attempt to present an overview of various representations with respect to language modelling, from neural word embeddings such as Word2Vec and GloVe to deep contextualized pre-trained embedding such as ULMFit, ELMo, OpenAI GPT and BERT. [ABSTRACT FROM AUTHOR]

Abstract This research paper presents a new approach that constitutes a first step towards programming using language-independent requirements. To leverage the needed programming effort, our approach takes requirements in the form of language-independent use case scenarios. Then, it generates the inputs of a code generator which, in turn, produces the corresponding code. To provide for the language-independence, our approach uses an enriched version of the semantic model, as a means to represent similar ideas possibly in different ways and in different natural languages. The enrichment consists of a set of patterns that it implements as XML code representing the information embedded in the use case scenarios. This intermediate representation can be processed to derive the inputs required by any code generator to produce code in a particular programming language. This paper illustrates the approach and its tool support for use case scenarios written in English and French, and semantic model patterns implemented as XML code that can be processed by the ReDSeeDS code generator. In addition, it presents the results of an experimental evaluation of the approach on use case scenarios (written in English and in French) belonging to five different systems. This evaluation quantitatively shows the ability of our approach: to extract ReDSeeDS inputs conforming to the expert's inputs with a high precision; to generate XML code elements conforming to the input with an encouraging performance as evaluated by the participating students (an F-measure ranging between 87.43% and 92.31%); and to generate Java code judged efficient by the participating programmers (an F-measure ranging between 66.4% and 93.43%). [ABSTRACT FROM AUTHOR]

The study of Online Social Networks offers growing opportunities to examine a number of aspects of the real world and to better understand how human society works at scale. One crucial research direction, constantly evolving over time, concerns the language adopted by users and its impact on online communities. To provide a contribution in this setting, in this paper, we adopt a multi-relational model for social network and identify a new typology of relation, namely co-interest, as an explicit common declaration of engagement towards a given topic among pair of users. Thematic communities can be, hence, derived leveraging such relations. After that, exploiting Natural Language Processing and Machine Learning techniques, we identify and define suitable analysis metrics focusing on the characteristics of the language adopted in such communities. With the objective of analyzing how the adoption of different language may impact the formation of cohesive and strong communities, to make our study general, we consider two very popular and intrinsically different social platforms, namely Twitter and Reddit. The analysis carried out in this paper aims at comparing the two social media to find interesting and novel results about how their users compose textual comments and the corresponding impact on users' interaction. Interestingly, the obtained results show that, although the main factor is related to the popularity of the user generating the content, the characteristics of the adopted language play a non negligible role in the formation of strong communities in social systems. • Investigation about language characteristics of Online Social Networks posts (OSN). • Language analysis to identify the evolution dynamics of online communities. • Correlation between potential post engagement and user adopted language style. • Relationship between the content produced in OSNs and the formation of communities. [ABSTRACT FROM AUTHOR]

Machine translation has become the most challenging as well as demanding field of research and as a result, to develop a fully automatic 100% correct machine translation system has not been achieved till date. More than 7000 of languages are being used worldwide for communication purposes and in India alone 22 languages have been recognized as official languages for communication purposes. Various efforts have been made to solve this problem by different researchers using different approaches. With the enhancement in technology and internet, automatic translation has become more important to share the knowledge. The authors in this paper described a mechanism of translation using machines for one of the ancient Indian Language Sanskrit to the latest computer understandable language Universal Networking Language (UNL). The proposed system uses two databases: one for the analysis purpose consists of more than 300 rules and the other for the generation of UNL which consists of about 1500 rules. For the performance evaluation, the authors selected 500 Sanskrit sentences in such a way that covers maximum UNL relations from various sources. The sentences are manually converted from English to Sanskrit language taken from Spanish Language center, Hindi to English language taken from IIT Bombay Hindi–UNL expressions and NCERT class 8th Sanskrit book titled "Ruchita Part 3". The proposed system achieves a BLEU score of 0.85. The proposed system gives 93.18% efficiency in resolving UNL relation successfully. [ABSTRACT FROM AUTHOR]

This paper presents a combination of R packages—user contributed toolkits written in a common core programming language—to facilitate the humanistic investigation of digitised, text-based corpora. Our survey of text analysis packages includes those of our own creation (cleanNLP and fasttextM) as well as packages built by other research groups (stringi, readtext, hyphenatr, quanteda, and hunspell). By operating on generic object types, these packages unite research innovations in corpus linguistics, natural language processing, machine learning, statistics, and digital humanities. We begin by extrapolating on the theoretical benefits of R as an elaborate gluing language for bringing together several areas of expertise and compare it to linguistic concordancers and other tool-based approaches to text analysis in the digital humanities. We then showcase the practical benefits of an ecosystem by illustrating how R packages have been integrated into a digital humanities project. Throughout, the focus is on moving beyond the bag-of-words, lexical frequency model by incorporating linguistically-driven analyses in research. [ABSTRACT FROM AUTHOR]

Natural Language Processing (NLP) is a field of computer science and linguistics concerned with the unique conversation between computers and human languages. It processes data through Lexical analysis, Syntax analysis, Semantic analysis, Discourse processing and Pragmatic analysis. An intelligent text summarization is one of the most challenging tasks in Natural language processing. It can be further used for applications like storytelling and question answering. This paper presents an automatic text summarizer for text documents using soft computing approach, consisting of SVO (Subject, Verb, and Object) Rules and Tag based training. This approach processes data through POS Tagger, NLP Parser, ambiguity removal, Semantic Representation, Sentence Reduction and Sentence Combination. At first, this paper defines the theme (title) of the document. After this operation, it preprocesses text document to perform pronominal reference resolution and text clustering. After these preprocessing operations, it identifies and removes ambiguity from the language using parser. And then, it calculates the score for the sentences using the title of the document, Semantic Sentence Similarity utility and n-gram Co-Occurrence relations of the words in a particular sentence. At last, sentences are combined with the SVO Rules after providing tag based training for simple and complex sentences. The summarizer was tested on the standard DUC 2007 dataset as well as a corpus of hundred text documents of different domains created by us. DUC 2007 Update Task produced accuracy F-scores of 0.13523 (ROUGE-2) and 0.112561 (ROUGE-SU4) for DUC 2007 documents and 0.4036 (ROUGE-2) and 0.3129 (ROUGE-SU4) for our corpus. Subjective evaluation was carried out by five language experts and twenty random individuals for system generated sample summaries. [ABSTRACT FROM AUTHOR]

FusionBrain is a research project aimed at the development of efficient multitask and multimodal models and their application to a wide variety of practical tasks. The general purpose and idea of the project is to learn to create models that can effectively extract additional important knowledge from a large number of data modalities and training tasks and, as a result, can better solve other tasks. The research is performed in many modalities: texts, images, audio, video, programming languages, graphs (e.g., molecular structures), time series, and so on. The lists of tasks to be solved is large and ranges from classical tasks in computer vision and natural language processing to tasks involving different modalities: VideoQA, Visual Commonsense Reasoning, and IQ tests (which are difficult to solve even for humans). The ability of models to solve tasks formulated in natural or visual languages and to cope with hidden tasks (for which there were no examples in the training set). Among other things, the studies focus on reduction in data and human and computational resources necessary at the training and inference stages. Some results concerning the study and development of multimodal and multitask architectures are described in this paper. [ABSTRACT FROM AUTHOR]

One of the main components of the Mizar project is the Mizar language, a computer language invented to reflect the natural language of mathematics. From the very beginning various linguistic constructions and grammar rules which enable us to write texts which resemble classical mathematical papers have been developed and implemented in the language. The Mizar Mathematical Library is a repository of computer-verified mathematical texts written in the Mizar language. Besides well-known and important theorems, the library contains series of some quite technical lemmas describing some properties formulated for different values of numbers. For example the sequence of lemmas for n being Nat st n < =1 holds n=0 or n=1; for n being Nat st n < =2 holds n=0 or n=1 or n=2; for n being Nat st n < =3 holds n=0 or n=1 or n=2 or n=3; which for a long time contained 13 such formulae. In this paper, we present an extension of the Mizar language – an ellipsis that is used to define flexary logical connectives. We define flexary conjunction and flexary disjunction, which can be understood as generalization of classical conjunction and classical disjunction, respectively. The proposed extension enables us to get rid of lists of such lemmas and to formulate them as single theorems, e.g. for m,n being Nat st n < =m holds n=0 or ... or n=m; covering all cases between the bounds 0 and m in this case. Moreover, a specific inference rule to process flexary formulae, formulae using flexary connectives, is introduced. We describe how ellipses are incorporated into the Mizar language and how they are verified by the Mizar proof checker. [ABSTRACT FROM AUTHOR]

• A novel manually annotated Alzheimer's disease dataset for low resource language i.e., Nepalese, consisting of 168 Alzheimer's disease and 98 control normal patients is presented. The dataset is publicly available for research community;. • This paper presented a NLP based framework for the early detection of AD patients using a Nepali transcripts and develop a visualization of content present in textual data. • This paper demonstrated a word cloud of most common words are presented to give qualitative analysis. • The performance of different state-of-the-art machine learning-based textual classification mechanisms are presented, and baseline results for each are reported. Alzheimer's disease (AD) is considered as progressing brain disease, which can be slowed down with the early detection and proper treatment by identifying the early symptoms. Language change serves as an early sign that a patient's cognitive functions have been impacted, potentially leading to early detection. The effects of language changes are being studied thoroughly in the English language to analyze the linguistic patterns in AD patients using Natural Language Processing (NLP). However, it has not been much explored in local languages and low-resourced languages like Nepali. In this paper, we have created a novel dataset on low resources language, i.e., Nepali, consisting of transcripts of the AD patients and control normal subjects. We have also presented baselines by applying various machine learning (ML) and deep learning (DL) algorithms on a novel dataset for the early detection of AD. The proposed work incorporates the speech decline of AD patients in order to classify them as control subjects or AD patients. This study makes an effective conclusion that the difficulty in processing information of AD patients reflects in their speech narratives of patients while describing a picture. The dataset is made publicly available. [ABSTRACT FROM AUTHOR]

Knowledge bases (KBs) such as Freebase and Yago are rather incomplete, and the situation is more serious in non-English KBs, such as Chinese KBs. In this paper, we present a language-independent framework to tackle the slot-filling task by searching the Web with high-precision queries, and deriving lightweight extraction patterns. The patterns are based on string matching, and since they make no use of complex NLP resources, which may be unavailable in some languages, they are very language-independent. We use a traditional bootstrapping approach for extraction, but also use a novel approach to suppress the noise associated with distant supervision: in particular, we use a pseudo-testing method to validate the patterns derived from different sentences. Experiments show that our framework achieves very encouraging results. [ABSTRACT FROM AUTHOR]

The analysis of time helps in extracting knowledge from web contents. In order to analyze massive amounts of data, current natural language processing systems rely mainly on supervised approaches: machine learning algorithms that learn how to classify data based on corpora, annotated with the TimeML mark-up language or one of its derivatives. The quality of annotation data, thus, affects the performances of such systems. Quality can be improved by reasoning on temporal annotations. This article takes such a view. After a review of the strictly necessary background, it focuses on and discusses open issues in the area of quality of temporal annotations: inconsistency and incompleteness of annotations. Then it proposes a semantic reasoning approach as solution for improving on their quality, viz., the SOA-based Qualitative Temporal Reasoner for reasoning about temporal annotations, which leverages on existing theories and tools for qualitative reasoning. The article presents the design of the reasoner and its two main reasoning services: consistency checking for tackling inconsistency, and deduction for addressing incompleteness on demand. Afterward, the paper presents an experimental evaluation of the reasoner, sustaining why the chosen semantic reasoning approach can help improve on quality of annotations. The experiment assesses the reasoner’s performances on two different corpora and from several perspectives, e.g., the effectiveness of consistency checking in terms of the number of inconsistent documents found, and of deduction in terms of the number of annotations added. It concludes with discussion of the results of the evaluation and possible routes for future work. [ABSTRACT FROM PUBLISHER]

GUI is an essential factor influencing users' perception of an app. As the visual bridge between the user and the application, the quality of a GUI should be measured by the users' subjective feelings. Therefore, to improve the UI, developers should pay attention to what users think and actively respond to their feedback. User reviews from different channels provide a way for developers to obtain users' suggestions about the UI, and it could be of great benefit to improve the UI of their apps. Methods to mine different information (e.g., bug report, feature request) from reviews have been proposed in the previous works, however, suggestion information for supporting UI improvement has not gained much attention. In this paper, we propose UISMiner (UI Suggestion Miner), a novel approach to automatically mine the suggestions about the UI design from reviews. The application of UISMiner relies on natural language processing and machine learning classifier, which involves three main steps: 1) gaining reviews related to UI suggestion by training a classifier; 2) extracting UI suggestions from the obtained reviews by defining rules. 3) associating UI suggestions with the corresponding part in the UI. The experiment based on Google Play shows that UISMiner can effectively obtain reviews related to UI suggestions and extract such information from the gained reviews, the F-measure can reach 87.71% and 76.99% respectively. Furthermore, the survey on 10 developers shows that the information provided by us is useful for updating the UI of apps. [ABSTRACT FROM AUTHOR]

This paper addresses the opportunities and challenges involved in applying modern approach. A model to realize a natural language interface for process control is described. The interface accepts natural language input and it is translated to a form that can activate concerned process controller to start up motorized valve while monitoring associated sensors .The interface can be reconfigured to suit changes in process without altering program thus eliminating the tiring task of reprogramming. [ABSTRACT FROM AUTHOR]

• A probabilistic model is proposed to evaluate the rating profiles of users. • A reliable mechanism is proposed to enhance the users' rating profiles. • A deep sparse autoencoder is used to extract latent features of users. • An integration approach is introduced to make recommendations. Deep neural networks have been extensively employed in many applications such as natural language processing and computer vision. They have attracted a lot of attention in designing recommender systems due to their recent significant achievements. These models are mainly used in recommender systems to generate the users' latent features according to the input data such as user-item rating and trust matrixes. However, such input data are significantly sparse leading to reduce the effectiveness of deep learning models in generating the latent features. In addition, the existing deep learning based recommendation models ignore the reliability measure in their recommendation process. To address these challenges, in this paper, a trust-aware recommendation method based on deep sparse autoencoder is proposed. Specifically, an effective probabilistic model is proposed to determine how many ratings are required for each user to produce an accurate prediction. Moreover, for enhancing the rating profiles of users with insufficient ratings, an efficient mechanism is developed based on the utilization of implicit ratings. In this mechanism, the implicit ratings are selected based on a reliability measure to guarantee their effectiveness. Then, the reliable enhanced rating profiles and trust statements are considered as the input data of deep sparse autoencoder to generate the users' latent features. Finally, a similarity function that utilizes the generated latent features is developed to obtain similarity values between users and make recommendations. Massive experiments are carried out on two datasets to assess the effectiveness of the proposed method in comparison to the-state-of-the-art models. The results of experiments demonstrate that the proposed method can significantly beat other models in producing more accurate recommendations. [ABSTRACT FROM AUTHOR]

This paper illustrates some of the knowledge representation structures and inference procedures proper to a high-level, fully implemented conceptual language, NKRL (Narrative Knowledge Representation Language). The aim is to show how these tools can be used to deal, in a sentiment analysis/opinion mining context, with some common types of human (and non-human) "behaviors". These behaviors correspond, in particular, to the concrete, mutual relationships among human and non-human characters that can be expressed under the form of non-fictional and real-time "narratives" (i.e., as logically and temporally structured sequences of "elementary events"). [ABSTRACT FROM AUTHOR]

Transformers have been widely utilized in recent NLP studies. Unlike CNNs or RNNs, the vanilla Transformer is position-insensitive, and thus is incapable of capturing the structural priors between sequences of words. Existing studies commonly apply one single mask strategy on Transformers for incorporating structural priors while failing at modeling more abundant structural information of texts. In this paper, we aim at introducing multiple types of structural priors into Transformers, proposing the Multiple Structural Priors Guided Transformer (MS-Transformer) that transforms different structural priors into different attention heads by using a novel multi-mask based multi-head attention mechanism. In particular, we integrate two categories of structural priors, including the sequential order and the relative position of words. For the purpose of capturing the latent hierarchical structure of the texts, we extract these information not only from the word contexts but also from the dependency syntax trees. Experimental results on three tasks show that MS-Transformer achieves significant improvements against other strong baselines. • Multi-mask strategies can introduce different priors into different attention heads. • Multi-mask strategies can guide models learning more precise dependencies. • The sequential order and relative position of words are taken as structure priors. • Structure priors benefit models on modeling sentence structures from multiple aspects. [ABSTRACT FROM AUTHOR]

The comprehension of source code is very difficult, especially if the programmer is not familiar with the programming language. Pseudocode explains and describes code contents that are based on the semantic analysis and understanding of the source code. In this paper, a novel retrieval-based transformer pseudocode generation model is proposed. The proposed model adopts different retrieval similarity methods and neural machine translation to generate pseudocode. The proposed model handles words of low frequency and words that do not exist in the training dataset. It consists of three steps. First, we retrieve the sentences that are similar to the input sentence using different similarity methods. Second, pass the source code retrieved (input retrieved) to the deep learning model based on the transformer to generate the pseudocode retrieved. Third, the replacement process is performed to obtain the target pseudo code. The proposed model is evaluated using Django and SPoC datasets. The experiments show promising performance results compared to other language models of machine translation. It reaches 61.96 and 50.28 in terms of BLEU performance measures for Django and SPoC, respectively. [ABSTRACT FROM AUTHOR]

Abstract: LTAG is a rich formalism for performing NLP tasks such as semantic interpretation, parsing, machine translation and information retrieval. Depend on the specific NLP task, different kinds of LTAGs for a language may be developed. Each of these LTAGs is enriched with some specific features such as semantic representation and statistical information that make them suitable to be used in that task. The distribution of these capabilities among the LTAGs makes it difficult to get the benefit from all of them in NLP applications. This paper discusses a statistical model to bridge between two kinds LTAGs for a natural language in order to benefit from the capabilities of both kinds. To do so, an HMM was trained that links an elementary tree sequence of a source LTAG onto an elementary tree sequence of a target LTAG. Training was performed by using the standard HMM training algorithm called Baum–Welch. To lead the training algorithm to a better solution, the initial state of the HMM was also trained by a novel EM-based semi-supervised bootstrapping algorithm. The model was tested on two English LTAGs, XTAG (XTAG-Group, 2001) and MICA''s grammar (Bangalore et al., 2009) as the target and source LTAGs, respectively. The empirical results confirm that the model can provide a satisfactory way for linking these LTAGs to share their capabilities together. [Copyright &y& Elsevier]

In conventional algorithms, the lack of entity information, reference, and semantic relations in the current corpus leads to a low rate of precision and efficiency in constructing cross‐language bilingual mapping. According to natural language processing and machine translation technology, to solve the problem, this paper aims to establish a parallel corpus for information extraction based on the OntoNotes corpus by combining automatic extraction and manual adjustment. To verify the validity of the parallel corpus constructed in this paper, a comparative experiment was carried out on the corpus. The corpus entity alignment rate, anaphora absence, and syntactic structure were analysed in detail based on statistics. The data set is well performed in language processing and machine translation. The parallel corpus for information extraction constructed in this paper can produce highly precise, stable, and efficient information in the process of bilingual mapping, which provides an effective parallel corpus for the study in machine translation of bilingual mapping. [ABSTRACT FROM AUTHOR]

[Display omitted] • An NLP transformer for relationship prediction between look-ahead tasks and master schedule activities with 82.1% Precision. • A generative model for creating look-ahead planning task descriptions based on a master schedule activity description. • A semi-automated human-in-the-loop method for linking look-ahead tasks to master schedule activities with 76.5% Precision. • A formalized representation for matching look-ahead planning tasks to master schedule activities. In construction, master schedules and look-ahead plans are created at different times (monthly vs. weekly), by different personas (planner vs. superintendent), with different software (scheduling solution vs. spreadsheet), and at different levels of granularity (milestones vs. production details). Their full-alignment is essential for project coordination, progress updating, and payment application reviews, and its absence may lead to costly litigation. This paper presents the first attempt to automate linking look-ahead planning tasks to master-schedule activities following an NLP-based multi-stage ranking formulation. Our model employs distance-based matching for candidate generation and a Transformer architecture for final matching. 1 1 Access to the code and a sample of the dataset can be granted upon direct request from the authors. Validation results from real-world projects demonstrate that the method helps planners match look-ahead planning tasks to master schedule activities by presenting a list of top-five matches with a precision of 76.5%. We also show that the method helps superintendents create look-ahead plans from a master schedule by generating lists of tasks based on activity descriptions. [ABSTRACT FROM AUTHOR]

Abstract: Written text is an important component in the process of knowledge acquisition and communication. Poorly written text fails to deliver clear ideas to the reader no matter how revolutionary and ground-breaking these ideas are. Providing text with good writing style is essential to transfer ideas smoothly. While we have sophisticated tools to check for stylistic problems in program code, we do not apply the same techniques for written text. In this paper we present TextLint, a rule-based tool to check for common style errors in natural language. TextLintprovides a structural model of written text and an extensible rule-based checking mechanism. [Copyright &y& Elsevier]

Objective: This paper examines the use of Semantic MEDLINE, a natural language processing application enhanced with a statistical algorithm known as Combo, as a potential decision support tool for clinicians. Semantic MEDLINE summarizes text in PubMed citations, transforming it into compact declarations that are filtered according to a user's information need that can be displayed in a graphic interface. Integration of the Combo algorithm enables Semantic MEDLINE to deliver information salient to many diverse needs. Methods: The authors selected three disease topics and crafted PubMed search queries to retrieve citations addressing the prevention of these diseases. They then processed the citations with Semantic MEDLINE, with the Combo algorithm enhancement. To evaluate the results, they constructed a reference standard for each disease topic consisting of preventive interventions recommended by a commercial decision support tool. Results: Semantic MEDLINE with Combo produced an average recall of 79% in primary and secondary analyses, an average precision of 45%, and a final average F-score of 0.57. Conclusion: This new approach to point-of-care information delivery holds promise as a decision support tool for clinicians. Health sciences libraries could implement such technologies to deliver tailored information to their users. [ABSTRACT FROM AUTHOR]

This paper proposes a new UGC-oriented language technology application, which we call experience mining. Experience mining aims at automatically collecting instances of personal experiences as well as opinions from vast amounts of user generated content (UGC) such as weblog and forum posts and storing them in an experience database with semantically rich indices. After discussing the technical issues relating to this new task, we focus on the central problem of factuality analysis, formulate a task definition, and propose a machine learning-based solution. Our empirical evaluation indicates that our factuality analysis defintion is sufficiently well-defined to achieve a high inter-annotator agreement and our Factorial CRF-based model considerably outperforms the baseline. We also present an application system, which currently stores over 50M experience instances extracted from 150M Japanese blog posts with semantic indices and serves an experience search engine for unrestricted users and report on our empirical evaluation of the system's accuracy. [ABSTRACT FROM AUTHOR]

This paper presents a formal syntax framework of natural languages for computational linguistics. The abstract syntax of natural languages, particularly English, and their formal manipulations are described. On the basis of the abstract syntax, a universal language processing model and the deductive grammar of English are developed toward the formalization of Chomsky's universal grammar in linguistics. Comparative analyses of natural and programming languages, as well as the linguistic perception on software engineering, are discussed. A wide range of applications of the deductive grammar of English have been explored in language acquisition, comprehension, generation, and processing in cognitive informatics, computational intelligence, and cognitive computing. [ABSTRACT FROM AUTHOR]

In Natural Language Processing (NLP) domain, a lexical database, i.e. a systematic stock of words of a natural language, is a crucial component of a wide variety of NLP applications. In this paper, we present the interlingua used in the construction of the bilingual lexical database called REBECA, one of the few lexical resources that encompasses Brazilian Portuguese (BP) language. The interlingua is the core feature of REBECA and it is composed of a set of concepts which allows the alignment of the English and BP monolingual databases. Precisely, we present: (i) the composition; (ii) the structure, (iv) the formalism, and (v) the editing tool used to develop REBECA. . As a result, we obtained a formal and hierarchical interlingua, which allowed a proper link between the monolingual databases. [ABSTRACT FROM AUTHOR]

Probabilistic language models have gained popularity in Natural Language Processing due to their ability to successfully capture language structures and constraints with computational efficiency. Probabilistic language models are flexible and easily adapted to language changes over time as well as to some new languages. Probabilistic language models can be trained and their accuracy strongly related to the availability of large text corpora. In this paper, we investigate the usability of grapheme probabilistic models, specifically grapheme n-grams models in spellchecking as well as augmentative typing systems. Grapheme n-gram models require substantially smaller training corpora and that is one of the main drivers for this thesis in which we build grapheme n-gram language models for the Albanian language. There are presently no available Albanian language corpora to be used for probabilistic language modeling. Our technique attempts to augment spellchecking and typing systems by utilizing grapheme n-gram language models in improving suggestion accuracy in spellchecking and augmentative typing systems. Our technique can be implemented in a standalone tool or incorporated in another tool to offer additional selection/scoring criteria. [ABSTRACT FROM AUTHOR]

Abstract: Natural language requirements specifications form the basis for the subsequent phase of the information system development process, namely the development of conceptual schemata. Both, the textual as well as the conceptual representations are not really appropriate for being thoroughly captured and validated by the ‘requirement holders’, i.e. the end users. Therefore, in our approach the textual specifications are firstly linguistically analyzed and translated into a so-called conceptual predesign schema. That schema is formulated using an interlingua which is based on a lean semantic model, thus allowing users to participate more efficiently in the design and validation process. After validation, the predesign schema is mapped to a conceptual representation (e.g. UML). The sequence of these translation and transformation steps is described by the “NIBA workflow”. This paper focuses on the information supporting a step by step mapping of natural language requirements specifications to a conceptual model, and on how that information is gained. On particular, we present a four-level interpretation of tagging-output. [Copyright &y& Elsevier]

Incomplete, inaccurate, ambiguous, and vola-tile requirements have plagued the software industry since its inception. The convergence of model-based development and formal methods offers developers of safety-critical systems a powerful new approach to the early validation of requirements. This paper describes an exercise conducted to determine if formal methods could be used to validate system requirements early in the lifecycle at reasonable cost. Several hundred functional and safety requirements for the mode logic of a typical flight guidance system were captured as natural language “shall” statements. A formal model of the mode logic was written in the RSML− e language and translated into the NuSMV model checker and the PVS theorem prover using translators developed as part of the project. Each “shall” statement was manually translated into a NuSMV or PVS property and proven using these tools. Numerous errors were found in both the original requirements and the RSML− e model. This demonstrates that formal models can be written for realistic systems and that formal analysis tools have matured to the point where they can be effectively used to find errors before implementation. [ABSTRACT FROM AUTHOR]

In this paper is presented the algorithm of an automatic grapheme-to-phoneme translator for the Greek language. The proposed algorithm is designed to collaborate with a high quality Text-to-Speech synthesis system in Greek. The algorithm assimilates the full reading process of written text as realized by a Greek speaking person. A detailed study of the Greek language's operation, led us to the implementation of an automatic integrated system which describes its phonetic behaviour in an exact and natural way. The software that implements the algorithm has the capability to receive written text from any input (keyboard, file, screen reader, OCR system e.t.c.) and transform it to phonetic form. Afterwards the output of the algorithm is directed to the input of a concatenation-based speech synthesizer and the right pronunciation of any written text is achieved in real-time. During the reading process the software locates and distinguishes Greek written text from any foreign language words, specially written symbols, abbreviations e. t. c... and afterwards manages them in order that the flow of the reading process permits the right perception of the produced spoken messages. The most important qualification of the algorithm is the possibility to incorporate it in other Text-to-Speech synthesis systems of different technology. Finally experimental measurements indicate the successful operation of the algorithm. [ABSTRACT FROM AUTHOR]

• Transfer fine-tuning yields representations suitable for specific tasks; in this paper we focused on sentence pair modelling. • The method helps the BERT model converge more quickly with a smaller corpus. • It also realizes performance gains while maintaining the model size. • Simple features outperform elaborate ones in phrasal paraphrase classification. Sentence pair modelling is defined as the task of identifying the semantic interaction between a sentence pair, i.e., paraphrase and textual entailment identification and semantic similarity measurement. It constitutes a set of crucial tasks for research in the area of natural language understanding. Sentence representation learning is a fundamental technology for sentence pair modelling, where the development of the BERT model realised a breakthrough. We have recently proposed transfer fine-tuning using phrasal paraphrases to allow BERT's representations to be suitable for semantic equivalence assessment between sentences while maintaining the model size. Herein, we reveal that transfer fine-tuning with simplified feature generation allows us to generate representations that are widely effective across different types of sentence pair modelling tasks. Detailed analysis confirms that our transfer fine-tuning helps the BERT model converge more quickly with a smaller corpus for fine-tuning. [ABSTRACT FROM AUTHOR]

Although a large number of natural language database interfaces have been developed, there have been few empirical studies of their practical usefulness. This paper presents the design and results of a field evaluation of a natural language system-NLS-used for data retrieval. A balanced multifactorial design comparing NLS to a reference retrieval language SQL is described. The data are analyzed on two levels: work task (n=87) and query (n=1081). SQL performed better than NLS on a variety of measures, but NLS required less effort to use. Subjects performed much poorer than expected based on the results of laboratory studies. This finding is attributed to the complexity of the field setting and to optimism in grading laboratory experiments. The methodology developed for studying computer languages in real work settings was successful in consistently measuring differences in treatments over a variety of conditions. [ABSTRACT FROM AUTHOR]

• RNNLG framework for language generation. • Heuristically gated LSTM generator. • Semantically controlled LSTM generator. • Attentive encoder decoder generator. • Data counterfeiting. • Discriminative training. • Corpus-based evaluation in a single domain setting. • Human evaluation in a single domain setting. • Corpus-based evaluation for domain adaptation. • Human evaluation for domain adaptation. Natural Language Generation (NLG) is a critical component of spoken dialogue systems and it has a significant impact both on usability and perceived quality. Most existing NLG approaches in common use employ rules and heuristics and tend to generate rigid and stylised responses without the natural variation of human language. Moreover, these limitations also add significantly to development costs and make the delivery of cross-domain, cross-lingual dialogue systems especially complex and expensive. The first contribution of this paper is to present RNNLG, a Recurrent Neural Network (RNN)-based statistical natural language generator that can learn to generate utterances directly from dialogue act – utterance pairs without any predefined syntaxes or semantic alignments. The presentation includes a systematic comparison of the principal RNN-based NLG models available. The second contribution, is to test the scalability of the proposed system by adapting models from one domain to another. We show that by pairing RNN-based NLG models with a proposed data counterfeiting method and a discriminative objective function, a pre-trained model can be quickly adapted to different domains with only a few examples. All of the findings presented are supported by both corpus-based and human evaluations. [ABSTRACT FROM AUTHOR]

Abstract: Purpose: A large share of the information in electronic medical records (EMRs) consists of free-text compositions. From a computational point-of-view, the continuing prevalence of free-text entry is a major hindrance when the goal is to increase automation in EMRs. However, the efforts in developing standards for the structured representation of medical information have not proven to be a panacea. The information space of clinical medicine is very diverse and constantly evolving, making it challenging to develop standards for the domain. This paper reports a study aiming to increase automation in the EMR through the computational understanding of specific class of medical text in English, namely emergency department chief complaints. Methods: We apply domain-specific analytical modeling for the computational understanding of chief complaints. We evaluate the performance of this approach in the automatic classification of chief complaints, e.g., for use in automatic syndromic surveillance. Results: The evaluation in a multi-hospital setting showed that the presented algorithm was accurate in terms of classification correctness. Also, use of approximate matching in the algorithm to cope with typographic variance did not affect classification correctness while increasing classification completeness. [Copyright &y& Elsevier]