Search

Multilingual pre-trained language models have achieved impressive results on most natural language processing tasks. However, the performance is inhibited due to capacity limitations and their under-representation of pre-training data, especially for languages with limited resources. This has led to the creation of tailored pre-trained language models, in which the models are pre-trained on large amounts of monolingual data or domain specific corpus. Nevertheless, compared to relying on multiple monolingual models, utilizing multilingual models offers the advantage of multilinguality, such as generalization on cross-lingual resources. To combine the advantages of both multilingual and monolingual models, we propose KDDA - a framework that leverages monolingual models to a single multilingual model with the aim to improve sentence representation for Vietnamese. KDDA employs teacher-student framework and cross-lingual transfer that aims to adopt knowledge from two monolingual models (teachers) and transfers them into a unified multilingual model (student). Since the representations from the teachers and the student lie on disparate semantic spaces, we measure discrepancy between their distributions by using Sinkhorn Divergence - an optimal transport distance. We conduct experiments on two Vietnamese natural language understanding tasks, including machine reading comprehension and natural language inference. Experimental results show that our model outperforms other state-of-the-art models and yields competitive performances. [ABSTRACT FROM AUTHOR]

Sports news is a type of discourse that is characterized by a specific vocabulary, style, and tone, and it is typically focused on conveying information about sporting events, athletes, and teams. Thematic context-based deep learning is a powerful approach that can be used to analyze and interpret various forms of natural language, including the discourse expression of sports news. An application model of sign language and lip language recognition based on deep learning is proposed to facilitate people with hearing impairment to easily obtain sports news content. First, the lip language recognition system is constructed; next, MobileNet lightweight network combined with Long-Short Term Memory (LSTM) is used to extract lip reading features. ResNet-50 residual network structure isadopted to extract the features of sign language; finally, the convergence, accuracy, precision and recall of the model are verified respectively. The results show that the loss of training set and test set converges gradually with the increase of iteration times; the lip language recognition model and the sign language recognition model basically tend to be stable after 14 iterations and 12 iterations, respectively, suggesting a better convergence effect of sign language recognition. The accuracy of sign language recognition and lip language recognition is 98.9% and 87.7%, respectively. In sign language recognition, the recognition accuracy of numbers 1, 2, 4, 6 and 8 can reach 100%. In lip language recognition, the recognition accuracy of numbers 2, 3 and 9 is relatively higher. This exploration can facilitate hearing-impaired people to quickly obtain the relevant content in sports news videos, and also provide help for their communication. [ABSTRACT FROM AUTHOR]

In recent years, contrastive learning has been very successful in unsupervised tasks of representation learning and has received a lot of attention in supervised tasks. In supervised tasks, the discrete nature of natural language makes the construction of sample pairs difficult and the models are poorly robust to adversarial samples, so it remains a challenge to make contrastive learning effective for text classification tasks and to guarantee the robustness of the models. This paper presents a contrastive adversarial learning framework built using data augmentation with labeled insertion data. Specifically,By adding perturbation to the word-embedding matrix, adversarial samples are generated as positive examples of contrastive learning, and external semantic information is introduced to construct negative examples. Contrastive learning is used to improve the sensitivity and generalization ability of the model, and adversarial training is used to improve robustness, thereby improving the classification accuracy. In addition, the momentum contrast from unsupervised tasks is also introduced into the text classification task to increase the number of sample pairs. Experimental results on several datasets show that the proposed approach outperforms the baseline comparison approach, and in addition some experiments are conducted to verify the effectiveness of the proposed framework under low-resource conditions. [ABSTRACT FROM AUTHOR]

The three-way decision model based on linguistic term sets has been extensively investigated since decision makers frequently utilize natural language to evaluate in an actual decision-making process. The existing models require decision makers to select appropriate linguistic terms from a given linguistic term set. However, making such a choice is not always simple, and decision makers occasionally choose words that are related to their own experience. In order to deal with this kind of decision problem, we appeal to the theory of computing with words pioneered by Zadeh and establish a three-way decision model based on computing with words in this paper. The paper focuses on how to deal with more general linguistic information using the theory of computing with words. Initially, using the concept of computing with words, we translate more broad linguistic information into a linguistic distribution assessment on a balanced linguistic term set in order to better analyze linguistic information. The three-way decision based on computing with words is then discussed. Decision-theoretic rough fuzzy sets take into account the ambiguity of the decision target as a generalization of the classical decision-theoretic rough sets. This is what motivated us to develop a three-way decision based on decision-theoretic rough fuzzy sets using computing with words. Additionally, a fabricated example demonstrates that our three-way decision model is more adaptable in processing linguistic information and can handle more general linguistic information provided by decision makers. [ABSTRACT FROM AUTHOR]

In this paper, we propose a method exploiting syntactic structure, semantic relations and word embeddings for recognizing textual entailment. The sentence pairs are analyzed using their syntactic structure and categorization of sentences in active voice, sentences in passive voice and sentences holding copular relations. The main syntactic relations such as subject, verb and object are extracted and lemmatized using a lemmatization algorithm based on parts-of-speech. The subject-to-subject, verb-to-verb and object-to-object similarity is identified using enhanced Wordnet semantic relations. Further similarity is analyzed using modifier relation, number relation, nominal modifier relation, compound relation, conjunction relation and negative relation. The experimental evaluation of the method on Stanford Natural Language Inference dataset shows that the accuracy of the method is 1.4% more when compared to the state-of-the-art zero shot domain adaptation methods. [ABSTRACT FROM AUTHOR]

Generating natural language description for visual content is a technique for describing the content available in the image(s). It requires knowledge of both the domains of computer vision and natural language processing. For this, various models with different approaches are suggested. One of them is encoder-decoder-based description generation. Existing papers used only objects for descriptions, but the relationship between them is equally essential, requiring context information. Which required techniques like Long Short-Term Memory (LSTM). This paper proposes an encoder-decoder-based methodology to generate human-like textual descriptions. Dense-LSTM is presented for better description as a decoder with a modified VGG19 encoder to capture information to describe the scene. Standard datasets Flickr8K and Flickr30k are used for testing and training purposes. BLEU (Bilingual Evaluation Understudy) score is used to evaluate the generated text. For the proposed model, a GUI (Graphical User Interface) is developed, which produces the audio description of the output received and provides an interface for searching the related visual content and query-based search. [ABSTRACT FROM AUTHOR]

Requirements are important in software development. Ambiguous requirements cause inconsistent understanding by developers, which leads to rework, delayed delivery, and other problems, and may even have devastating effects on the project. A large number of requirements text written in natural language are not concise, intuitive, and accurate. This condition increases the workload of designers and the difficulty of their tasks. An effective solution for the aforementioned problems is to extract actors and use cases from the requirement texts. This study proposes a model for extracting actors and using cases automatically, which combines bi-directional long short-term memory (BiLSTM) and conditional random fields. BiLSTM is used to capture the contextual information of the texts, and CRF is used to calculate the tag transfer score and determine the most accurate tag sequence, which aims to extract actors and use cases. Results show that the accuracy of extraction is significantly improved compared with the baseline method, which verifies the effectiveness of the proposed method in extracting actors and use cases. [ABSTRACT FROM AUTHOR]

Sentiment analysis is a task that belongs to natural language processing and it is highly used in texts extracted from social networks. This task consists of assigning the labels or classes: positive, negative or neutral to the text. However, analyzing a piece of text extracted from social networks to determine if it represents a positive or negative sentiment is a difficult task, because social media texts contain slangs, typographical errors and cultural context. The shortcomings of traditional frequency based feature extraction models such as bag of words or TF-IDF affect the accuracy of sentiment classification. To improve the precision in the sentiment classification task, it is possible to use natural language modelling methods that are able to learn contextual information from words. In this work, word embedding such as Word2Vec, GloVe and Doc2VecC with different dimensions are used. The resulting word vectors will be used to train recurring neural networks such as LSTM, BiLSTM, GRU and BiGRU, to improve sentiment classification. [ABSTRACT FROM AUTHOR]

Semantic merger, which is a phenomenon of semantic convergence of two meanings of a word in a certain context, is a kind of semantic indeterminacy in natural language, however, it brings trouble for natural language processing. Discovery of the features causing semantic mergers has been a significant but tough issue in natural language processing. Until now this issue has remained untouched. Therefore, in this article, this issue is studied. Based on a 1.8 million word English multi-genre corpus, taking English modal verb may as the target word, the contextual features causing semantic mergers between may(root possibility) and may(epistemic possibility) are investigated by an approach of attribute partial order diagram (APOD). First, the objects of may is categorized into 3 classes based on the idea of the three-way decision: may1(root possibility), may2(epistemic possibility) and may3(merger), then the rules for word sense disambiguation (WSD) of the three classes are extracted, respectively, and a comparison is made among the rules for different classes, and finally the features causing semantic merger of may are discovered. The discovered knowledge provides valuable evidence for finding the semantic merger, the cause of the semantic merger and the solution of semantic mergers of may, and the proposed approach can also be use for other modal verbs, which may benefit the natural language processing of English modal verbs. [ABSTRACT FROM AUTHOR]

Named Entity Recognition (NER) is a vitally important task of Natural Language Processing (NLP), which aims at finding named entities in natural language text and classifying them into predefined categories such as persons (PER), places (LOC), organizations (ORG), and so on. In the Arabic context, the current NER approaches based on deep learning are mainly based on word embedding or character-level embedding as input. However, using a single granularity representation has problems with out-of-vocabulary (OOV), word embedding errors, and relatively simple semantic content. This paper presents a multi-headed self-attention mechanism implemented in the BiLSTM-CRF neural network structure to recognize Arabic named entities on social media using two embeddings. Unlike other state-of-the-art approaches, this approach combines character and word embedding at the embedding layer, and the attention mechanism calculates the similarity over the entire sequence of characters and captures local context information. The proposed approach better recognized NEs in Dialect Arabic, reaching an F1 value of 74.15% on Darwish's dataset (a publicly available Arabic NER benchmark for social media). According to our knowledge, our findings outperform the current state-of-the-art models for Arabic Named Entity Recognition on social media. [ABSTRACT FROM AUTHOR]

The confluence of high performance computing algorithms and large scale high-quality data has led to the availability of cutting edge tools in computational linguistics. However, these state-of-the-art tools are available only for the major languages of the world. The preparation of large scale high-quality corpora for low-resource language such as Urdu is a challenging task as it requires huge computational and human resources. In this paper, we build and analyze a large scale Urdu language Twitter corpus Anbar. For this purpose, we collect 106.9 million Urdu tweets posted by 1.69 million users during one year (September 2018-August 2019). Our corpus consists of tweets with a rich vocabulary of 3.8 million unique tokens along with 58K hashtags and 62K URLs. Moreover, it contains 75.9 million (71.0%) retweets and 847K geotagged tweets. Furthermore, we examine Anbar using a variety of metrics like temporal frequency of tweets, vocabulary size, geo-location, user characteristics, and entities distribution. To the best of our knowledge, this is the largest repository of Urdu language tweets for the NLP research community which can be used for Natural Language Understanding (NLU), social analytics, and fake news detection. [ABSTRACT FROM AUTHOR]

Person re-identification with natural language description is a process of retrieving the corresponding person's image from an image dataset according to a text description of the person. The key challenge in this cross-modal task is to extract visual and text features and construct loss functions to achieve cross-modal matching between text and image. Firstly, we designed a two-branch network framework for person re-identification with natural language description. In this framework we include the following: a Bi-directional Long Short-Term Memory (Bi-LSTM) network is used to extract text features and a truncated attention mechanism is proposed to select the principal component of the text features; a MobileNet is used to extract image features. Secondly, we proposed a Cascade Loss Function (CLF), which includes cross-modal matching loss and single modal classification loss, both with relative entropy function, to fully exploit the identity-level information. The experimental results on the CUHK-PEDES dataset demonstrate that our method achieves better results in Top-5 and Top-10 than other current 10 state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

In the last decade, many intelligent interfaces and layers have been suggested to allow the use of relational databases and extraction of the content using only the natural language. However most of them struggle when exposed to new databases. In this article, we present SQLSketch, a sketch-based network for generating SQL queries to address the problem of automatically translate Natural Languages questions to SQL using the related databases schemas. We argue that the previous models that use full or partial sequence-to-sequence structure in the decoding phase can, in fact, have counter-effect on the generation operation and came up with more loss of the context or the meaning of the user question. In this regard, we use a full sketch-based structure that decouples the generation process into many small prediction modules. The SQLSketch is evaluated against GreatSQL, a new cross-domain, large-scale and balanced dataset for the Natural Language to SQL translation task. For a long-term aim of making better models and contributing in adding more improvements to the semantic parsing tasks, we propose the GreatSQL dataset as the first balanced cross-domain corpus that includes 45,969 pairs of natural language questions and their corresponding SQL queries in addition to simplified and well structured ground-truth annotations. We establish results for SQLSketch using GreatSQL dataset and compare the performance against two popular types of models that represent the sequential and partial-sketch based approaches. Experimental result shows that SQLSketch outperforms the baseline models by 13% in exact matching accuracy and achieve a score of 23.9% to be the new state-of-the-art model on GreatSQL. [ABSTRACT FROM AUTHOR]

In general, there are a lot of uncertainties in uncertain information, natural language, and human knowledge. The conclusion can be better deduced by using an approximate reasoning method, while a fuzzy intelligent system can deal with uncertain data and rule evaluation information systems. In order to better explore diffusion and economic growth, this paper constructs a fuzzy intelligent system based on the DSGE model and uses this system to analyze diffusion and economic growth. In order to verify the feasibility of this system, we test the response time and accuracy of the system. In addition, we also use the system to simulate diffusion and economic growth. The results show that with the increase of the task amount, the gap between the actual response time and the expected response time of the fuzzy intelligent system based on the DSGE model increases. When the task quantity is 20, the expected response time is 2.31 and the actual response time is 2.24. When the task quantity is 40, the expected response time is 2.5 and the actual response time is 2.36. The larger the task quantity is, the faster the response time of a fuzzy intelligent system based on the DSEG model is. Therefore, the fuzzy intelligent system based on the DSEG model has good performance and can analyze diffusion and economic growth well. [ABSTRACT FROM AUTHOR]

At present, the field of natural language will also introduce in-depth learning, using the concept of word vector, so that the neural network can also complete the work in the field of statistics. It can be said that the neural network has begun to show its advantages in the field of natural language processing. In this paper, the author analyzes the multimedia English course based on fuzzy statistics and neural network clustering. Different factors were classified, and scores were classified according to the number of characteristics of different categories. It can be seen that with the popularization of the Internet, MOOC teaching meets the requirements of the current college English curriculum, is a breakthrough in the traditional teaching mode, improves students' participation, and enables students to learn independently. It not only conforms to the characteristics of College students, but also improves their learning effect. In the automatic scoring stage, the quantitative text features are extracted by the feature extractor in the pre-processing stage, and then the weights of network connections obtained in the training stage are used to score the weights comprehensively. This model can better reflect students' autonomous learning ability and language application ability. [ABSTRACT FROM AUTHOR]

The corpus software has many functions, such as keyword retrieval, context co-occurrence, word list generation and word frequency statistics. It can quickly and accurately provide various corpus and information, such as word-formation collocation, context, word frequency and so on. In this paper, the author analyzes the application of deep learning and target visual detection in English vocabulary online teaching. Deep learning is a kind of machine learning algorithm which includes multi-layer non-linear mapping and tries to obtain high-level abstract representation of data. By extracting features from information, the identifiable components in the image can be extracted. The results show that the application of corpus in College English vocabulary teaching can promote students'autonomous use of corpus in English vocabulary learning. The simulation experiment improves the performance of the system by choosing parameters, and the classification accuracy is more than 90%. Corpus can enable students to learn real and natural language and master natural collocation. At the same time, corpus can help students understand the semantic and pragmatic norms of words in communication and recognize the characteristics of register variants. Future research can use Map-reduce technology to accelerate the training process, save training time and test more hyperparameters. [ABSTRACT FROM AUTHOR]

English speech recognition system is affected by a variety of interference factors. Associating the algorithm with the support of modern computer technology can increase the model effect of speech recognition system. Based on the study of the current mainstream controlled natural language thesaurus, this paper proposes a controlled natural language vocabulary classification type. Moreover, this paper defines the domain thesaurus according to the WordNet knowledge description framework, and uses WordNet's synonym, antisense, upper and lower, etc. In this way, the controlled natural language system can use the semantic relationship of WordNet to identify the words of the non-domain thesaurus input by the user and map the non-domain definition words to the words in the domain thesaurus, thereby improving the ease of use of controlled natural language systems. In addition, this paper designed a controlled experiment to analyze the performance of this system. The research results show that the model constructed in this paper has certain significant effects. [ABSTRACT FROM AUTHOR]

This paper presents a novel deep learning based approach to solving arithmetic word problems. Solving different types of mathematical (math) word problems (MWP) is a very complex and challenging task as it requires Natural Language Understanding (NLU) and Commonsense knowledge. An application on this can benefit learning (education) technologies such as E-learning systems, Intelligent tutoring, Learning Management Systems (LMS), Innovative teaching/learning, etc. We propose Deep Learning based Arithmetic Word Problem Solver, DLAWPS, an intelligent MWP solver system. DLAWPS consists of a Recurrent Neural Network (RNN) based Bi-directional Long Short-Term Memory (BiLSTM) to classify operation among four basic operations {+ , - , * , /}, and a knowledge-based irrelevant information removal unit (IIRU) to identify the relevant quantities to form an equation to solve arithmetic MWPs. Our system generates state-of-the-art results on the standard arithmetic word problem datasets –AddSub, SingleOp, and a Combined dataset. [ABSTRACT FROM AUTHOR]

In this work, we report the results of our experiments on the task of distinguishing the semantics of verb-noun collocations in a Spanish corpus. This semantics was represented by four lexical functions of the Meaning-Text Theory. Each lexical function specifies a certain universal semantic concept found in any natural language. Knowledge of collocation and its semantic content is important for natural language processing, as collocation comprises the restrictions on how words can be used together. We experimented with word2vec embeddings and six supervised machine learning methods most commonly used in a wide range of natural language processing tasks. Our objective was to study the ability of word2vec embeddings to represent the context of collocations in a way that could discriminate among lexical functions. A difference from previous work with word embeddings is that we trained word2vec on a lemmatized corpus after stopwords elimination, supposing that such vectors would capture a more accurate semantic characterization. The experiments were performed on a collection of 1,131 Excelsior newspaper issues. As the experimental results showed, word2vec representation of collocations outperformed the classical bag-of-words context representation implemented in a vector space model and fed into the same supervised learning methods. [ABSTRACT FROM AUTHOR]

With the development and popularization of electronic computers and the Internet, the problem of language barriers has once again become prominent in the new era, and people are more in need of machine translation. However, there is currently no suitable method for effective semantic ordering of English machine translation. In order to better perform semantic ordering on English machine translation, the article combines fuzzy theory to construct an algorithm model, and analyzes the experimental results through evaluation indicators. The results show that with the increase of training concentration training examples, the semantic parser can learn more natural language sentence analysis methods from the training examples, and the natural language sentences that can be correctly parsed gradually increase, so with the training examples increased recall rate and F value gradually increased. The experimental results also show that the use of higher precision syntax analyzers can effectively improve the performance of statistical machine translation systems, whether in phrase-based or machine-based translation methods. [ABSTRACT FROM AUTHOR]

The basic system of inquisitive semantics (InqB) established by Groenendijk et al. is a general inquisitive semantic theory which doesn't concern fuzziness. To explain the fuzzy phenomena in natural languages, this paper extends InqB into the framework of M-fuzzifying setting and establishes a basic system of M-fuzzifying inquisitive semantics. To begin with, the notion of M-fuzzifying supporting mapping is defined, where M is a completely distributive lattice with an involution operator and each subset of the universal set of all possible worlds can be regarded as a support of any well-formed formula to some degree. Then the notions of M-fuzzifying entailment order, M-fuzzifying truth mappings, M-fuzzifying informative content mappings and M-fuzzifying inquisitive content mappings are introduced and their properties are discussed. Further, the degrees of assertiveness, informativeness, inquisitiveness and questioning of a well-formed formula are defined, by which the M-fuzzifying assertive projection operator and the M-fuzzifying questioning projection operator are introduced and characterized. Finally, a necessary and sufficient condition is obtained, where a well-formed formula is exactly the disjunction of its unique M-fuzzifying assertive projection and unique M-fuzzifying questioning projection. [ABSTRACT FROM AUTHOR]

Information Extraction allows machines to decipher natural language through using two tasks: Named Entity Recognition and Relation Extraction. In order to build such a system for Bangla Language, in this work a Named Entity Recognition (NER) System is proposed, which requires a minimum information to deliver a decent performance having less dependency on handcrafted features. The proposed model is based on Deep Learning, which is accomplished through the use of a Densely Connected Network (DCN) in collaboration with a Bidirectional-LSTM (BiLSTM) and word embedding, i.e., DCN-BiLSTM. Such a system, specific to the Bangla language, has never been done before. Furthermore, a unique dataset was made since no Named Entity Recognition dataset exists for Bangla language till date. In the dataset, over 71 thousand Bangla sentences have been collected, annotated, and classified into four different groups using IOB tagging scheme. Those groups are person, location, organization, and object. Due to Bangla's morphological structure, character level feature extraction is also applied so that we can access more features to determine relational structure between different words. This is initially done with the use of a Convolutional Neural Network but is later outperformed by our second approach which is through the use of a Densely Connected Network (DCN). As for the training portion, it has been done for two variations of word embedding which are word2vec and glove, the outcome being the largest vocabulary size known to both models. A detailed discussion in regard to the methodology of the NER system is explained in a comprehensive manner followed by an examination of the various evaluation scores achieved. The proposed model in this work resulted in having a F1 score of 63.37, which is evaluated at Named Entity Level. [ABSTRACT FROM AUTHOR]

In spite of having been investigated for over fifty years, developing a robust spoken dialog management system remains an open research issue in robotics and natural language processing. In this paper, we present a language-independent spoken dialog management module integrated into a human-robot interaction system. We adopt an algorithmic approach to dialog modeling. A mobile robot functioning as a shopping assistant exemplifies the proposed approach. The dialog module is composed of a state transition network, in which state switches are conditioned by both visual and communicative factors. We use the formalism of a finite state automaton, where the robot changes its state by performing a speech act or a non-verbal action from the set of specified act/action types. [ABSTRACT FROM AUTHOR]

Detection of topics in Natural Language text collections is an important step towards flexible automated text handling, for tasks like text translation, summarization, etc. In the current dominant paradigm to topic modeling, topics are represented as probability distributions of terms. Although such models are theoretically sound, their high computational complexity makes them difficult to use in very large scale collections. In this work we propose an alternative topic modeling paradigm based on a simpler representation of topics as overlapping clusters of semantically similar documents, that is able to take advantage of highly-scalable clustering algorithms. Our Query-based Topic Modeling framework (QTM) is an information-theoretic method that assumes the existence of a "golden" set of queries that can capture most of the semantic information of the collection and produce models with maximum "semantic coherence". QTM was designed with scalability in mind and was executed in parallel using a Map-Reduce implementation; further, we show complexity measures that support our scalability claims. Our experiments show that the QTM can produce models of comparable or even superior quality than those produced by state of the art probabilistic methods. [ABSTRACT FROM AUTHOR]

Question Answering Systems (QASs) are search engines that have the ability to provide a brief and accurate answer to each question in natural languages. The question asked in such a system is answered with a set of documents, a paragraph, a sentence, etc. In this paper, a solution is proposed to optimize the performance of web-based QASs for answering definitional and factoid questions in English. As evolutionary algorithms are suitable for issues with large search space and also texts can be examined from a variety of aspects, this approach proposes for the first time employing Multi-Objective Evolutionary algorithms (MOEAs) to optimize the performance of QASs. In the present work, we provided a Multi-Objective QAS (MOQAS) that would be more accurate in choosing the most probable answer from the documents that the standard search engine has retrieved. Through the ranking process, various features can be extracted from the text. Each of these features examines the text from a perspective, but changes in the values of these features are not consistent with each other; thus, we need to use a method that takes all these views into account. For this purpose, three different aspects of the text are considered and NSGA-II as a MOEA is applied. We used two standard datasets and web data for evaluating our system. Comparing the obtained results from the proposed MOQAS with other existing QASs reveals that MOQAS yields promising and effective results. [ABSTRACT FROM AUTHOR]

Machine Translation (MT) requires a large amount of linguistic resources, which leads current MT systems to leaving unknownwords untranslated. This can be annoying for end users, as they might not understand at all such untranslated words. However, most language families share a common vocabulary, therefore this knowledge can be leveraged to produce more understandable translations, typically for "assimilation" or gisting use. Based on this observation, we propose a method that constructs polyglot translations tailored to a particular user language. Simply put, an unknown word is translated into a set of languages that relate to the user's language, and the translated word that is closest to the user's language is used as a replacement of the unknown word. Experimental results on language coverage over three language families indicate that our method may improve the usefulness of MT systems. As confirmed by a subsequent human evaluation, polyglot translations look indeed familiar to the users, and are perceived to be easier to read and understand than translations in their related natural languages. [ABSTRACT FROM AUTHOR]

In this paper we introduce an objective method for CVSS score calculation. CVSS is a well known and mostly used method for giving priority to software vulnerabilities. Currently it is being calculated by some slightly subjective methods which require enough skill and knowledge. This research shows how we can benefit from natural language description of vulnerabilities for CVSS calculation. The data that were used for implementation and evaluation of the proposed models consists of the available CVE vulnerability descriptions and their corresponding CVSS scores from the OSVDB database. First, feature vectors were extracted using text mining tools and techniques, and then the SVM and Random-Forest algorithms as well as fuzzy systems were examined to predict the concerned CVSS scores. In spite of the fact that SVM and Random-Forest are mostly used and trusted methods in prediction, results of this research bear a witness that using fuzzy systems can give comparable and even better results. In addition, implementation of the fuzzy based system is much easier and faster. Although so far, there have been so little efforts in using the information embedded in textual materials regarding vulnerabilities, this research shows that it will be valuable to utilize them in systems security establishment. [ABSTRACT FROM AUTHOR]