Search results for: google word2vec word embeddings

Authors: Sanja Seljan, Ivan Dunđer

Abstract:

The paper presents combined automatic speech recognition (ASR) for English and machine translation (MT) for English and Croatian in the domain of business correspondence. The first part presents results of training the ASR commercial system on two English data sets, enriched by error analysis. The second part presents results of machine translation performed by online tool Google Translate for English and Croatian and Croatian-English language pairs. Human evaluation in terms of usability is conducted and internal consistency calculated by Cronbach's alpha coefficient, enriched by error analysis. Automatic evaluation is performed by WER (Word Error Rate) and PER (Position-independent word Error Rate) metrics, followed by investigation of Pearson’s correlation with human evaluation.

Keywords: automatic machine translation, integrated language technologies, quality evaluation, speech recognition

Procedia PDF Downloads 452

Authors: Shelby Simpson, William Stanley, Namir Naba, Xiaodi Wang

Abstract:

Both t-SNE and UMAP are brand new state of art tools to predominantly preserve the local structure that is to group neighboring data points together, which indeed provides a very informative visualization of heterogeneity in our data. In this research, we develop a t-SNE and UMAP base neural network image classification algorithm to embed the original dataset to a corresponding low dimensional dataset as a preprocessing step, then use this embedded database as input to our specially designed neural network classifier for image classification. We use the fashion MNIST data set, which is a labeled data set of images of clothing objects in our experiments. t-SNE and UMAP are used for dimensionality reduction of the data set and thus produce low dimensional embeddings. Furthermore, we use the embeddings from t-SNE and UMAP to feed into two neural networks. The accuracy of the models from the two neural networks is then compared to a dense neural network that does not use embedding as an input to show which model can classify the images of clothing objects more accurately.

Keywords: t-SNE, UMAP, fashion MNIST, neural networks

Procedia PDF Downloads 164

Authors: Usama Shahid

Abstract:

Be it written documents, news columns, or students' essays, verifying the content can be a time-consuming task. Apart from the spelling and grammar mistakes, the proofreader is also supposed to verify whether the content included in the essay or document is relevant or not. The irrelevant content in any document or essay is referred to as off-topic text and in this paper, we will address the problem of off-topic text detection from a document using machine learning techniques. Our study aims to identify the off-topic content from a document using Echo state network model and we will also compare data with other models. The previous study uses Convolutional Neural Networks and TFIDF to detect off-topic text. We will rearrange the existing datasets and take new classifiers along with new word embeddings and implement them on existing and new datasets in order to compare the results with the previously existing CNN model.

Keywords: off topic, text detection, eco state network, machine learning

Procedia PDF Downloads 53

Authors: Hussah M. AlShayea

Abstract:

This study aimed to identify the potential of cloud computing, "Google Drive" in overcoming the problems of collective learning from the viewpoint of Princess Noura University students. The study included (92) students from the College of Education. To achieve the goal of the study, several steps have been taken. First, the most important problems of collective learning were identified from the viewpoint of the students. After that, a survey identifying the potential of cloud computing "Google Drive" in overcoming the problems of collective learning was distributed among the students. The study results showed that the students believe that the use of Google Drive contributed to overcoming these problems. In the light of those results, the researcher presented a set of recommendations and proposals, including: encouraging teachers and learners to employ cloud computing to overcome the problems and constraints of collective learning.

Keywords: cloud computing, collective learning, Google drive, Princess Noura University

Procedia PDF Downloads 450

Authors: Jude K. Safo

Abstract:

Knowledge Graphs KG) and their relation to Graph Embeddings (GE) represent a unique data structure in the landscape of machine learning (relative to image, text and acoustic data). Unlike the latter, GEs are the only data structure sufficient for representing hierarchically dense, semantic information needed for use-cases like supply chain data and protein folding where the search space exceeds the limits traditional search methods (e.g. page-rank, Dijkstra, etc.). While GEs are effective for compressing low rank tensor data, at scale, they begin to introduce a new problem of ’data retreival’ which we observe in Large Language Models. Notable attempts by transE, TransR and other prominent industry standards have shown a peak performance just north of 57% on WN18 and FB15K benchmarks, insufficient practical industry applications. They’re also limited, in scope, to next node/link predictions. Traditional linear methods like Tucker, CP, PARAFAC and CANDECOMP quickly hit memory limits on tensors exceeding 6.4 million nodes. This paper outlines a topological framework for linear mapping between concepts in KG space and GE space that preserve cardinality. Most importantly we introduce a traceable framework for composing dense linguistic strcutures. We demonstrate performance on WN18 benchmark this model hits. This model does not rely on Large Langauge Models (LLM) though the applications are certainy relevant here as well.

Keywords: representation theory, large language models, graph embeddings, applied algebraic topology, applied knot theory, combinatorics

Procedia PDF Downloads 41

Authors: André Felipe Gimenez, Flávia Alessandra Ribeiro da Silva, Roberto Saverio Souza Costa

Abstract:

Currently, the rural producer has been facing problems regarding environmental regularization, which is precisely why the CAR (Rural Environmental Registry) was created. CAR is an electronic registry for rural properties with the purpose of assimilating notions about legal reserve areas, permanent preservation areas, areas of limited use, stable areas, forests and remnants of native vegetation, and all rural properties in Brazil. . The objective of this work was to evaluate and compare altimetry and slope data from google Earth with a digital terrain model (MDT) generated by aerophotogrammetry, in three plots of a steep slope, for the purpose of declaration in the CAR (Rural Environmental Registry). The realization of this work is justified in these areas, in which rural landowners have doubts about the reliability of the use of the free software Google Earth to diagnose inclinations greater than 25 degrees, as recommended by federal law 12651/2012. Added to the fact that in the literature, there is a deficiency of this type of study for the purpose of declaration of the CAR. The results showed that when comparing the drone altimetry data with the Google Earth image data, in areas of high slope (above 40% slope), Google underestimated the real values of terrain slope. Thus, it is concluded that Google Earth is not reliable for diagnosing areas with an inclination greater than 25 degrees (46% declivity) for the purpose of declaration in the CAR, being essential to carry out the local topographic survey.

Keywords: MDT, drone, RPA, SiCar, photogrammetry

Procedia PDF Downloads 96

Authors: Kevin Chan, Kevin K. H. Chung, Patcy P. S. Yeung, H. L. Ip, Bill T. C. Chung, Karen M. K. Chung

Abstract:

Purpose: Recent evidence shows that reading self-efficacy and students perceived difficulty in reading are significantly associated with word reading and reading fluency. However, little is known about these relationships among students learning to read English as a second language, particularly in Chinese students. This study examined the contributions of reading self-efficacy, perception of difficulty in reading, and cognitive-linguistic skills to performance on English word reading and reading fluency in Chinese students. Method: A sample of 122 second-and third-grade students in Hong Kong, China, participated in this study. Students completed the measures of reading self-efficacy and perception of difficulty in reading. They were assessed on their English cognitive-linguistic and reading skills: rapid automatized naming, nonword reading, phonological awareness, word reading, and one-minute word reading. Results: Results of path analysis indicated that when students’ grades were controlled, reading self-efficacy was a significant correlate of word reading and reading fluency, whereas perception of difficulty in reading negatively predicted word reading. Conclusion: These findings underscore the importance of taking students’ reading self-efficacy and perception of difficulty in reading and their cognitive-linguistic skills into consideration when designing reading intervention and instructions for students learning English as a second language.

Keywords: self-efficacy, perception of difficulty in reading, english as a second language, word reading

Procedia PDF Downloads 165

Authors: Sean Goltz, Michael Mayo

Abstract:

The purpose of this research is to use Global-Regulation.com database of the world laws, focusing on tax treaties between countries, in order to create an AI-driven tax simulator that will run an AI agent through potential tax scenarios across countries. The AI agent goal is to identify the scenario that will result in minimum tax liability based on tax treaties between countries. The results will be visualized by a three dimensional matrix. This will be an online web application. Multinational corporations are running their business through multiple countries. These countries, in turn, have a tax treaty with many other countries to regulate the payment of taxes on income that is transferred between these countries. As a result, planning the best tax scenario across multiple countries and numerous tax treaties is almost impossible. This research propose to use Global-Regulation.com database of word laws in English (machine translated by Google and Microsoft API’s) in order to create a simulator that will include the information in the tax treaties. Once ready, an AI agent will be sent through the simulator to identify the scenario that will result in minimum tax liability. Identifying the best tax scenario across countries may save multinational corporations, like Google, billions of dollars annually. Given the nature of the raw data and the domain of taxes (i.e., numbers), this is a promising ground to employ artificial intelligence towards a practical and beneficial purpose.

Keywords: taxation, law, multinational, corporation

Procedia PDF Downloads 166

Authors: S. B. Rathna Kumar, Pranjali A Ujwane, Panchanan Mohanty

Abstract:

The present study aimed to develop a battery for assessing speech recognition performance by adults in Marathi. A total of four word lists were developed by considering word frequency, word familiarity, words in common use, and phonemic balance. Each word list consists of 25 words (15 monosyllabic words in CVC structure and 10 monosyllabic words in CVCV structure). Equivalence analysis and performance-intensity function testing was carried using the four word lists on a total of 150 native speakers of Marathi belonging to different regions of Maharashtra (Vidarbha, Marathwada, Khandesh and Northern Maharashtra, Pune, and Konkan). The subjects were further equally divided into five groups based on above mentioned regions. It was found that there was no significant difference (p > 0.05) in the speech recognition performance between groups for each word list and between word lists for each group. Hence, the four word lists developed were equally difficult for all the groups and can be used interchangeably. The performance-intensity (PI) function curve showed semi-linear function, and the groups’ mean slope of the linear portions of the curve indicated an average linear slope of 4.64%, 4.73%, 4.68%, and 4.85% increase in word recognition score per dB for list 1, list 2, list 3 and list 4 respectively. Although, there is no data available on speech recognition tests for adults in Marathi, most of the findings of the study are in line with the findings of research reports on other languages. The four word lists, thus developed, were found to have sufficient reliability and validity in assessing speech recognition performance by adults in Marathi.

Keywords: speech recognition performance, phonemic balance, equivalence analysis, performance-intensity function testing, reliability, validity

Procedia PDF Downloads 328

Authors: Armineh Nourbakhsh, Sameena Shah, Carolyn Rose

Abstract:

Layout-aware language models have been used to create multimodal representations for documents that are in image form, achieving relatively high accuracy in document understanding tasks. However, the large number of parameters in the resulting models makes building and using them prohibitive without access to high-performing processing units with large memory capacity. We propose an alternative approach that can create efficient representations without the need for a neural visual backbone. This leads to an 80% reduction in the number of parameters compared to the smallest SOTA model, widely expanding applicability. In addition, our layout embeddings are pre-trained on spatial and visual cues alone and only fused with text embeddings in downstream tasks, which can facilitate applicability to low-resource of multi-lingual domains. Despite using 2.5% of training data, we show competitive performance on two form understanding tasks: semantic labeling and link prediction.

Keywords: layout understanding, form understanding, multimodal document understanding, bias-augmented attention

Procedia PDF Downloads 115

Authors: Mst Shapna Akter, Hossain Shahriar

Abstract:

One of the most important challenges in the field of software code audit is the presence of vulnerabilities in software source code. Every year, more and more software flaws are found, either internally in proprietary code or revealed publicly. These flaws are highly likely exploited and lead to system compromise, data leakage, or denial of service. C and C++ open-source code are now available in order to create a largescale, machine-learning system for function-level vulnerability identification. We assembled a sizable dataset of millions of opensource functions that point to potential exploits. We developed an efficient and scalable vulnerability detection method based on deep neural network models that learn features extracted from the source codes. The source code is first converted into a minimal intermediate representation to remove the pointless components and shorten the dependency. Moreover, we keep the semantic and syntactic information using state-of-the-art word embedding algorithms such as glove and fastText. The embedded vectors are subsequently fed into deep learning networks such as LSTM, BilSTM, LSTM-Autoencoder, word2vec, BERT, and GPT-2 to classify the possible vulnerabilities. Furthermore, we proposed a neural network model which can overcome issues associated with traditional neural networks. Evaluation metrics such as f1 score, precision, recall, accuracy, and total execution time have been used to measure the performance. We made a comparative analysis between results derived from features containing a minimal text representation and semantic and syntactic information. We found that all of the deep learning models provide comparatively higher accuracy when we use semantic and syntactic information as the features but require higher execution time as the word embedding the algorithm puts on a bit of complexity to the overall system.

Keywords: cyber security, vulnerability detection, neural networks, feature extraction

Procedia PDF Downloads 45

Authors: Pardeep Singh, Kamlesh Dutta

Abstract:

Many languages have fixed sentence structure and others are free word order. The accuracy of anaphora resolution of syntax based algorithm depends on structure of the sentence. So, it is important to analyze the structure of any language before implementing these algorithms. In this study, we analyzed the sentence structure exploiting the case marker in Hindi as well as some special tag for subject and object. We also investigated the word order for Hindi. Word order typology refers to the study of the order of the syntactic constituents of a language. We analyzed 165 news items of Ranchi Express from EMILEE corpus of plain text. It consisted of 1745 sentences. Eight file of dialogue based from the same corpus has been analyzed which will have 1521 sentences. The percentages of subject object verb structure (SOV) and object subject verb (OSV) are 66.90 and 33.10, respectively.

Keywords: anaphora resolution, free word order languages, SOV, OSV

Procedia PDF Downloads 442

Authors: Pelumi Olowofoyeku

Abstract:

English as a language has great reach and influence; it is taught all over the world. For instance, in Nigeria, English language is been taught and learned as a second language; therefore second learners of English in Nigeria have certain problems they contend with. Because of the dissimilarities in word formation patterns of English and Yoruba languages, Yoruba learners of English mostly found in the south west of Nigeria, and some parts of Kwara, Kogi, and Edo states of Nigeria have problems with word formation patterns in English. The objectives of this paper therefore, are: to identify the levels of word formation dissimilarities in English and Yoruba languages and to examine the effects of these dissimilarities on the Yoruba learners of English. The data for this paper were graded words purposely selected and presented to selected students of Adeniran Ogunsanya College of Education, Oto-Ijanikin, Lagos, who are Yoruba learners of English. These respondents were randomly selected to form words which are purposively selected to test the effects of word formation dissimilarities between Yoruba (the respondent’s first language) and English language on the respondents. The dissimilarities are examined using contrastive analysis tools. This paper reveals that there are differences in the word formation patterns of Yoruba and English languages. The writer believes that there is need for language teachers to undertake comparative studies of the two languages involved for methodological reasons. The author then suggests that teachers should identify the problem areas and systematically teach their students. The paper concludes that although English and Yoruba word formation patterns differ very significantly in many respects, there exist language universals in all languages which language educators should take advantage of in teaching.

Keywords: word formation patterns, graded words, ESL, Yoruba learners

Procedia PDF Downloads 468

Authors: Pui-sze Yeung, Connie Suk-han Ho, David Wai-ock Chan, Kevin Kien-hoa Chung

Abstract:

Background: Recent findings have shown that transcription skills play a unique and significant role in Chinese word reading and spelling (i.e. word dictation), and written composition development. The interrelationships among component skills of transcription, word reading, word spelling, and written composition in Chinese have rarely been examined in the literature. Is the contribution of component skills of transcription to Chinese written composition mediated by word level skills (i.e., word reading and spelling)? Methods: The participants in the study were 249 Chinese children in Grade 1, Grade 3, and Grade 5 in Hong Kong. They were administered measures of general reasoning ability, orthographic knowledge, stroke sequence knowledge, word spelling, handwriting fluency, word reading, and Chinese narrative writing. Orthographic knowledge- orthographic knowledge was assessed by a task modeled after the lexical decision subtest of the Hong Kong Test of Specific Learning Difficulties in Reading and Writing (HKT-SpLD). Stroke sequence knowledge: The participants’ performance in producing legitimate stroke sequences was measured by a stroke sequence knowledge task. Handwriting fluency- Handwriting fluency was assessed by a task modeled after the Chinese Handwriting Speed Test. Word spelling: The stimuli of the word spelling task consist of fourteen two-character Chinese words. Word reading: The stimuli of the word reading task consist of 120 two-character Chinese words. Written composition: A narrative writing task was used to assess the participants’ text writing skills. Results: Analysis of covariance results showed that there were significant between-grade differences in the performance of word reading, word spelling, handwriting fluency, and written composition. Preliminary hierarchical multiple regression analysis results showed that orthographic knowledge, word spelling, and handwriting fluency were unique predictors of Chinese written composition even after controlling for age, IQ, and word reading. The interaction effects between grade and each of these three skills (orthographic knowledge, word spelling, and handwriting fluency) were not significant. Path analysis results showed that orthographic knowledge contributed to written composition both directly and indirectly through word spelling, while handwriting fluency contributed to written composition directly and indirectly through both word reading and spelling. Stroke sequence knowledge only contributed to written composition indirectly through word spelling. Conclusions: Preliminary hierarchical regression results were consistent with previous findings about the significant role of transcription skills in Chinese word reading, spelling and written composition development. The fact that orthographic knowledge contributed both directly and indirectly to written composition through word reading and spelling may reflect the impact of the script-sound-meaning convergence of Chinese characters on the composing process. The significant contribution of word spelling and handwriting fluency to Chinese written composition across elementary grades highlighted the difficulty in attaining automaticity of transcription skills in Chinese, which limits the working memory resources available for other composing processes.

Keywords: orthographic knowledge, transcription skills, word reading, writing

Procedia PDF Downloads 390

Authors: D. Pugazhendhi

Abstract:

The ancient Greeks were forerunners in many fields than other societies. So, the Greeks were well connected with all the countries which were well developed during that time through trade route. In this connection, trading of goods from the ancient Greece to Tamil Nadu which is presently in India, though they are geographically far away, played an important role. In that way, the word and the goods related with kos and kare got exchanged between these two societies. So, it is necessary to compare the phonology and the morphological occurrences of these words that are found common both in the ancient Greek and Tamil literatures of the contemporary period. The results show that there were many words derived from the root kos with the basic meaning of ‘arrange’ in the ancient Greek language, but this is not the case in the usage of the word kare. In the ancient Tamil literature, the word ‘kos’ does not have any root and also had rare occurrences. But it was just the opposite in the case of the word ‘kare’. One of all the meanings of the word, which was derived from the root ‘kos’ in ancient Greek literature, is related with costly ornaments. This meaning seems to have close resemblance with the usage of word ‘kos’ in ancient Tamil literature. Also, the meaning of the word ‘kare’ in ancient Tamil literature is related with spices whereas, in the ancient Greek literature, its meaning is related to that of the cooking of meat using spices. Hence, the similarity seen in the meanings of these words ‘kos’ and ‘kare’ in both these languages provides lead for further study. More than that, the ancient literary resources which are available in both these languages ensure the export and import of gold and spices from the ancient Greek land to Tamil land.

Keywords: arrange, kare, Kos, ornament, Tamil

Procedia PDF Downloads 104

Authors: David N. Prihoda

Abstract:

This paper charts the history of the English word Tempest from its origins in Proto-Indo European to its modern usage as a term for storms, both literal and metaphorical. It does so by way of considering the word’s morphology, semiotics, and phonetics. It references numerous language studies and dictionaries to chronicle the word’s many steps along that path, from demarcation of measurement to assessment of time, all the way to an observation about the weather or the human psyche. The conclusive findings show that tempest has undergone numerous changes throughout its history, and these changes interestingly parallel its connotations as a symbol for both chaotic weather and the chaos of the human spirit

Keywords: Tempest, etymology, language origins, English

Procedia PDF Downloads 79

Authors: Vishal Raj, Noorhan Abbas

Abstract:

Ambiguity in NLP (Natural language processing) refers to the ability of a word, phrase, sentence, or text to have multiple meanings. This results in various kinds of ambiguities such as lexical, syntactic, semantic, anaphoric and referential am-biguities. This study is focused mainly on solving the issue of Lexical ambiguity. Word Sense Disambiguation (WSD) is an NLP technique that aims to resolve lexical ambiguity by determining the correct meaning of a word within a given context. Most WSD solutions rely on words for training and testing, but we have used lemma and Part of Speech (POS) tokens of words for training and testing. Lemma adds generality and POS adds properties of word into token. We have designed a novel method to create an affinity matrix to calculate the affinity be-tween any pair of lemma_POS (a token where lemma and POS of word are joined by underscore) of given training set. Additionally, we have devised an al-gorithm to create the sense clusters of tokens using affinity matrix under hierar-chy of POS of lemma. Furthermore, three different mechanisms to predict the sense of target word using the affinity/similarity value are devised. Each contex-tual token contributes to the sense of target word with some value and whichever sense gets higher value becomes the sense of target word. So, contextual tokens play a key role in creating sense clusters and predicting the sense of target word, hence, the model is named Contextual SenSe Model (CSM). CSM exhibits a noteworthy simplicity and explication lucidity in contrast to contemporary deep learning models characterized by intricacy, time-intensive processes, and chal-lenging explication. CSM is trained on SemCor training data and evaluated on SemEval test dataset. The results indicate that despite the naivety of the method, it achieves promising results when compared to the Most Frequent Sense (MFS) model.

Keywords: word sense disambiguation (wsd), contextual sense model (csm), most frequent sense (mfs), part of speech (pos), natural language processing (nlp), oov (out of vocabulary), lemma_pos (a token where lemma and pos of word are joined by underscore), information retrieval (ir), machine translation (mt)

Procedia PDF Downloads 66

Authors: Mohamed Ali Al Shaari, Khalid M. El Fitori

Abstract:

In the field of Quran Studies known as Ghareeb A Quran (the study of the meanings of strange words and structures in Holy Quran), it is difficult to distinguish some pragmatic meanings from conceptual meanings. One who wants to study this subject may need to look for a common usage between any two words or more; to understand general meaning, and sometimes may need to look for common differences between them, even if there are synonyms (word sisters). Some of the distinguished scholars of Arabic linguistics believe that there are no synonym words, they believe in varieties of meaning and multi-context usage. Based on this viewpoint, our method was designed to look for synonyms of a word, then the differences that distinct the word and their synonyms. There are many available books that use such a method e.g. synonyms books, dictionaries, glossaries, and some books on the interpretations of strange vocabulary of the Holy Quran, but it is difficult to look up words in these written works. For that reason, we proposed a logical entity, which we called Differences Matrix (DM). DM groups the synonyms words to extract the relations between them and to know the general meaning, which defines the skeleton of all word synonyms; this meaning is expressed by a word of its sisters. In Differences Matrix, we used the sisters(words) as titles for rows and columns, and in the obtained cells we tried to define the row title (word) by using column title (her sister), so the relations between sisters appear, the expected result is well defined groups of sisters for each word. We represented the obtained results formally, and used the defined groups as a base for building the ontology of the Holy Quran synonyms.

Keywords: Quran, synonyms, differences matrix, ontology

Procedia PDF Downloads 388

Authors: Seyedeh Khadijeh. Mirbazel, Masoumeh Arjmandi

Abstract:

Semantics is a linguistic approach and a scientific stream, and like all scientific streams, it is dynamic. The study of meaning is carried out in the broad semantic collections of words that form the discourse. In other words, meaning is not something that can be found in a word; rather, the formation of meaning is a process that takes place in a discourse as a whole. One of the contemporary semantic theories is Izutsu's Semantic Field Theory. According to this theory, the discovery of meaning depends on the function of words and takes place within the context of language. The purpose of this research is to identify the meaning of the word "Nas" in the discourse of verse 24 of Surah Al-Baqarah, which introduces "Nas" as the firewood of hell, but the translators have translated it as "people". The present research has investigated the semantic structure of the word "Nas" using the aforementioned theory through the descriptive-analytical method. In the process of investigation, by matching the semantic fields of the Quranic word "Nas", this research came to the conclusion that "Nas" implies those persons who have forgotten God and His covenant in believing in His Oneness. For this reason, God called them "Nas (the forgetful)" - the imperfect participle of the noun /næsiwoɔn/ in single trinity of Arabic language, which means “to forget”. Therefore, the intended meaning of "Nas" in the verses that have the word "Nas" is not equivalent to "People" which is a general noun.

Keywords: Nas, people, semantics, semantic field theory.

Procedia PDF Downloads 160

Authors: Xiuying Wang, Changliang Li, Bo Xu

Abstract:

A fixed-length feature vector is required for many machine learning algorithms in NLP field. Word embeddings have been very successful at learning lexical information. However, they cannot capture the compositional meaning of sentences, which prevents them from a deeper understanding of language. In this paper, we introduce a novel hierarchical tree long short-term memory (HTLSTM) model that learns vector representations for sentences of arbitrary syntactic type and length. We propose to split one sentence into three hierarchies: short phrase, long phrase and full sentence level. The HTLSTM model gives our algorithm the potential to fully consider the hierarchical information and long-term dependencies of language. We design the experiments on both English and Chinese corpus to evaluate our model on sentiment analysis task. And the results show that our model outperforms several existing state of the art approaches significantly.

Keywords: deep learning, hierarchical tree long short-term memory, sentence representation, sentiment analysis

Procedia PDF Downloads 328

Authors: Duygu Dere, Mert Ergeneci, Kaan Gokcesu

Abstract:

Google Trends data has gained increasing popularity in the applications of behavioral finance, decision science and risk management. Because of Google’s wide range of use, the Trends statistics provide significant information about the investor sentiment and intention, which can be used as decisive factors for corporate and risk management fields. However, an anomaly, a significant increase or decrease, in a certain query cannot be detected by the state of the art applications of computation due to the random baseline noise of the Trends data, which is modelled as an Additive white Gaussian noise (AWGN). Since through time, the baseline noise power shows a gradual change an adaptive thresholding method is required to track and learn the baseline noise for a correct classification. To this end, we introduce an online method to classify meaningful deviations in Google Trends data. Through extensive experiments, we demonstrate that our method can successfully classify various anomalies for plenty of different data.

Keywords: adaptive data processing, behavioral finance , convex optimization, online learning, soft minimum thresholding

Procedia PDF Downloads 134

Authors: Yahya Maeni

Abstract:

It is well established that amblyopic patients have a reduced reading performance and oculomotor deficits. Word length has a significant impact on reading performance and eye movement behaviour during reading. As there no previous attempts to assess whether amblyopic eyes would be affected by word length while reading. This study aims to assess the effect of word length on amblyopic eye movement behaviour during reading including fixation duration, number of fixation and gaze duration. 21 adults with amblyopia and 21 age-matched controls participated in the study (age ± SD) (23.80 ± 4.66) for amblyopes and (24.20 ± 3.58) for Controls. Eye movement was recorded during reading binocularly using Eyelink 1000. Study was designed as 2 x 2 (amblyopia vs. control) x 2 lengths (4 letters, and 8 letters). Compared to controls, the amblyopic participants report significant longer duration of fixation, higher number of fixation and longer gaze duration for short words with far higher significant difference for long words. It could be concluded that eye movement in amblyopia during reading might be accounted for by the length of a word within a text and this could possible explanation of reduced reading performance among amblyopes. By understanding the effect of word length on amblyopia will shed light on reading deficits in amblyopia and help to determine the reading needs of amplyopes in educational and clinical settings.

Keywords: amblyopia, eye movement, reading, fixation

Procedia PDF Downloads 118

Authors: Aris R. Terzopoulos, Georgia Z. Niolaki, Lynne G. Duncan, Mark A. J. Wilson, Antonios Kyparissiadis, Jackie Masterson

Abstract:

As word databases for bilingual children are not available, researchers, educators and textbook writers must rely on monolingual databases. The aim of this study is thus to develop a bilingual word database, BiLex-kids, an online open access developmental word database for 5-13 year old bilingual children who learn Greek as a second language and have English as their dominant one. BiLex-kids is compiled from 120 Greek textbooks used in Greek-English bilingual education in the UK, USA and Australia, and provides word translations in the two languages, pronunciations in Greek, and psycholinguistic variables (e.g. Zipf, Frequency per million, Dispersion, Contextual Diversity, Neighbourhood size). After clearing the textbooks of non-relevant items (e.g. punctuation), algorithms were applied to extract the psycholinguistic indices for all words. As well as one total lexicon, the database produces values for all ages (one lexicon for each age) and for three age bands (one lexicon per age band: 5-8, 9-11, 12-13 years). BiLex-kids provides researchers with accurate figures for a wide range of psycholinguistic variables, making it a useful and reliable research tool for selecting stimuli to examine lexical processing among bilingual children. In addition, it offers children the opportunity to study word spelling, learn translations and listen to pronunciations in their second language. It further benefits educators in selecting age-appropriate words for teaching reading and spelling, while special educational needs teachers will have a resource to control the content of word lists when designing interventions for bilinguals with literacy difficulties.

Keywords: bilingual children, psycholinguistics, vocabulary development, word databases

Procedia PDF Downloads 286

Authors: Saadia Elamin

Abstract:

The quasi-universal use of smart phones with internet connection available all the time makes it a reflex action for translation undergraduates, once they encounter the least translation problem, to turn to the freely available web resource: Google Translate. Like for other translator resources and aids, the use of Google Translate needs to be moderated in such a way that it contributes to developing translation competence. Here, instead of interfering with students’ learning by providing ready-made solutions which might not always fit into the contexts of use, it can help to consolidate the skills of analysis and transfer which students have already acquired. One way to do so is by training students to adhere to the basic principles of translation work. The most important of these is that analyzing the source text for comprehension comes first and foremost before jumping into the search for target language equivalents. Another basic principle is that certain translator aids and tools can be used for comprehension, while others are to be confined to the phase of re-expressing the meaning into the target language. The present paper reports on the experience of making a measured and reasonable use of Google Translate in translation teaching at Prince Sultan University (PSU), Riyadh. First, it traces the development that has taken place in the field of translation in this age of information technology, be it in translation teaching and translator training, or in the real-world practice of the profession. Second, it describes how, with the aim of reflecting this development onto the way translation is taught, senior students, after being trained on post-editing machine translation output, are authorized to use Google Translate in classwork and assignments. Third, the paper elaborates on the findings of this case study which has demonstrated that Google Translate, if used at the appropriate levels of training, can help to enhance students’ ability to perform different translation tasks. This help extends from the search for terms and expressions, to the tasks of drafting the target text, revising its content and finally editing it. In addition, using Google Translate in this way fosters a reflexive and critical attitude towards web resources in general, maximizing thus the benefit gained from them in preparing students to meet the requirements of the modern translation job market.

Keywords: Google Translate, post-editing machine translation output, principles of translation work, translation competence, translation teaching, translator aids and tools

Procedia PDF Downloads 441

Authors: Methawee Yuttapongtada

Abstract:

Intensifier is the linguistic term and device that is generally found in different languages in order to enhance and give additional quantity, quality or emotion to the words of each language. In fact, each language in the world has both of the similar and dissimilar intensifying device. More specially, the wide variety of intensifying device is used for Thai language and one of those is usage of the impolite word or the word that used to mean something negative as intensifier. The data collection in this study was done throughout the spoken language style by collecting from intensifiers regarded as impolite words because these words as employed in the other contexts will be held as the rude, swear words or the words with negative meaning. Then, backward study to the past was done in order to consider the historical change. Explanation of the original meaning and the contexts of words use from the past till the present time were done by use of both textual documents and dictionaries available in different periods. It was found that regarding the semantics and pragmatic aspects, subjectification also is the significant motivation that changed the impolite words to intensifiers. At last, it can explain pathway of the semantic change of these very words undoubtedly. Moreover, it is found that use tendency in the impolite word or the word that used to mean something negative will more be increased and this phenomenon is commonly found in many languages in the world and results of this research may support to the belief that human language in the world is universal and the same still reflected that human has the fundamental thought as the same to each other basically.

Keywords: impolite word, intensifier, Thai, semantic change

Procedia PDF Downloads 151

Authors: Jim Shih-Chiao Chin, Yu Ting Yeh, Shui Lien Chen, Yi-Fen Tsai

Abstract:

his study is taking Adidas Company as the object, explored the brand awareness directly or indirectly affects brand affect and word of mouth. First, explored the brand awareness on category fit and image fit, and examined the influence of category fit and image fit on extension attitude. This study then designates the effect of extension attitude on brand affect and word-of-mouth. The relationship of brand awareness on brand affect and word-of-mouth was also explored. The study participants are people who have purchased Adidas extension products. A total of 700 valid questionnaires were collected and statistical software AMOS 20.0 was used to examine the research hypotheses by using structural equation modeling (SEM). Finally, theoretical implications and research directions are provided for future studies.

Keywords: brand extension, brand awareness, product category fit, brand image fit, brand affect, word-of-mouth (WOM)

Procedia PDF Downloads 303

Authors: J. Marie Hicks

Abstract:

The art of the poetic conjoins meaning and symbolism with imagery and rhythm. Perhaps the reader might read this opening sentence as 'The art of the poetic combines meaning and symbolism with imagery and rhythm,' which holds a similar message, but is not quite the same. The reader understands that these factors are combined in this literary form, but to gain a sense of the conjoining of these factors, the reader is forced to consider that these aspects of poetry are not simply combined, but actually adjoin, abut, skirt, or touch in the poetic form. This alternative word choice is an example of substitutional inference. Poetry is, ostensibly, a literary form where language is used precisely or creatively to evoke specific images or emotions for the reader. Often, the reader can predict a coming rhyme or descriptive word choice in a poem, based on previous rhyming pattern or earlier imagery in the poem. However, there are instances when the poet uses an unexpected word choice to create multiple meanings and connections. In these cases, the reader is presented with an unusual phrase or image, requiring that they think about what that image is meant to suggest, and their mind also suggests the word they expected, creating a second, overlying image or meaning. This is what is meant by the term 'substitutional inference.' This is different than simply using a double entendre, a word or phrase that has two meanings, often one complementary and the other disparaging, or one that is innocuous and the other suggestive. In substitutional inference, the poet utilizes an unanticipated word that is either visually or phonetically similar to the expected word, provoking the reader to work to understand the poetic phrase as written, while unconsciously incorporating the meaning of the line as anticipated. In other words, by virtue of a word substitution, an inference of the logical word choice is imparted to the reader, while they are seeking to rationalize the word that was actually used. There is a substitutional inference of meaning created by the alternate word choice. For example, Louise Bogan, 4th Poet Laureate of the United States, used substitutional inference in the form of homonyms, malapropisms, and other unusual word choices in a number of her poems, lending depth and greater complexity, while actively engaging her readers intellectually with her poetry. Substitutional inference not only adds complexity to the potential interpretations of Bogan’s poetry, as well as the poetry of others, but provided a method for writers to infuse additional meanings into their work, thus expressing more information in a compact format. Additionally, this nuancing enriches the poetic experience for the reader, who can enjoy the poem superficially as written, or on a deeper level exploring gradations of meaning.

Keywords: poetic inference, poetic word play, substitutional inference, word substitution

Procedia PDF Downloads 201

Authors: Raghad Dwaik

Abstract:

It is well known in language acquisition literature that a mere understanding of a reading text is not enough to help students build proficiency in comprehension. Students should rather follow understanding by attempting to express what has been understood by pushing their competence to the limit. Learners' attempt to push their competence was given the term "comprehensible output" by Swain (1985). Teachers in large classes, however, find it sometimes difficult to give all students a chance to communicate their views or to share their ideas during the short class time. In most cases, students who are outgoing dominate class discussion and get more opportunities for practice which leads to ignoring the shy students totally while helping the good ones become better. This paper presents the idea of using Google Classroom features of posting and commenting to allow students who hesitate to participate in class discussions about a reading text to write their views on the wall of a Google Classroom and share them later after they have received feedback and comments from classmates. Such attempts lead to developing their proficiency through additional practice in comprehensible output and to enhancing their confidence in themselves and their views. It was found that virtual classroom interaction would help students maintain vocabulary, use more complex structures and focus on meaning besides form.

Keywords: learning groups, reading TESOL, Google Classroom, comprehensible output

Procedia PDF Downloads 38

Authors: Bok Check Meng, Goh Ying Soon

Abstract:

Giving additional learning materials such as Chinese fantasy novel to non-native learners can be strenuous. Instructors have to understand the underpinning theories about cognitive theory for new word instruction. This paper discusses the underpinning theories. Relevant literature reviews are given. There are basically five major areas of cognitive related theories mentioned in this article. These include motivational learning theory, Affective theory of learning, Cognitive psychology theory, Vocabulary acquisition theory and Bloom’s cognitive levels theory. A theoretical framework has been constructed. Thus, this will give a hand in ensuring non-native learners might gain positive outcomes in the instruction process. Instructors who are interested in teaching new word from Chinese fantasy novel in specific to support additional learning might be able to get insights from this article.

Keywords: Chinese fantasy novel, new word teaching, non-native learners, cognitive theory, bloom

Procedia PDF Downloads 702

Authors: Amna Asif, Rabia Naseem

Abstract:

In the domain of consumer-brand relationship, love for a fashion brand is a dominant idea. Brand executives incline to build more endearing brands, for example, Levi’s “Quality never goes out of style”. Though, the significance of this notion is not often debated in the literature of marketing. Moreover, the effect of brand image and personality on brand love has not been examined in any quantitative study in Pakistan. The current research aims to fill this study gap by evolving a causal framework integrating word-of-mouth, brand love, image, and personality to examine the relationships among them. Data was gathered through questionnaires survey, and it was filled by 409 university students. AMOS 20 was used to draw a path analysis and test the hypotheses. Results discovered that brand personality and brand image leads to brand love that ultimately impacts word-of-mouth. Results give thorough suggestions on which future research can be constructed.

Keywords: brand love, brand personality, brand image, fashion brands, word-of-mouth

Procedia PDF Downloads 280