Search results for: oral language competence

Authors: B. S. Premalatha, Mereen Rose Babu, Vaishali Prabhu

Abstract:

Background: Dysphagia must be assessed, either subjectively or objectively, in order to properly address the swallowing difficulty. Providing therapeutic care to patients with dysphagia via tele mode was one approach for providing clinical services during the COVID-19 epidemic. As a result, the teleassessment of dysphagia has increased in India. Aim: This study aimed to identify challenges faced by Indian SLPs while providing teleassessment to individuals with dysphagia during the outbreak of COVID-19 from 2020 to 2021. Method: After receiving approval from the institute's institutional review board and ethics committee, the current study was carried out. The study was cross-sectional in nature and lasted from 2020 to 2021. The study enrolled participants who met the inclusion and exclusion criteria of the study. It was decided to recruit roughly 246 people based on the sample size calculations. The research was done in three stages: questionnaire development and content validation, questionnaire administration. Five speech and hearing professionals' content verified the questionnaire for faults and clarity. Participants received questionnaires via various social media platforms such as e-mail and WhatsApp, which were written in Microsoft Word and then converted to Google Forms. SPSS software was used to examine the data. Results: In light of the obstacles that Indian SLPs encounter, the study's findings were examined. Only 135 people responded. During the COVID-19 lockdowns, 38% of participants said they did not deal with dysphagia patients. After the lockout, 70.4% of SLPs kept working with dysphagia patients, while 29.6% did not. From the beginning of the oromotor examination, the main problems in completing tele evaluation of dysphagia have been highlighted. Around 37.5% of SLPs said they don't undertake the OPME online because of difficulties doing the evaluation, such as the need for repeated instructions from patients and family members and trouble visualizing structures in various positions. The majority of SLPs' online assessments were inefficient and time-consuming. A bigger percentage of SLPs stated that they will not advocate tele evaluation in dysphagia to their colleagues. SLPs' use of dysphagia assessment has decreased as a result of the epidemic. When it came to the amount of food, the majority of people proposed a small amount. Apart from placing the patient for assessment and gaining less cooperation from the family, most SLPs found that Internet speed was a source of concern and a barrier. Hearing impairment and the presence of a tracheostomy in patients with dysphagia proved to be the most difficult conditions to treat online. For patients with NPO, the majority of SLPs did not advise tele-evaluation. In the anterior region of the oral cavity, oral meal residue was more visible. The majority of SLPs reported more anterior than posterior leakage. Even while the majority of SLPs could detect aspiration by coughing, many found it difficult to discern the gurgling tone of speech after swallowing. Conclusion: The current study sheds light on the difficulties that Indian SLPs experience when assessing dysphagia via tele mode, indicating that tele-assessment of dysphagia is still to gain importance in India.

Keywords: dysphagia, teleassessment, challenges, Indian SLP

Procedia PDF Downloads 141

Authors: L. Bruckert, A. Mansourzadeh

Abstract:

The voice has a major influence on interpersonal communication in everyday life via the perception of pleasantness. The evolutionary perspective postulates that the mechanisms underlying the pleasantness judgments are universal adaptations that have evolved in the service of choosing a mate (through the process of sexual selection). From this point of view, the favorite voices would be those with more marked sexually dimorphic characteristics; for example, in men with lower voice pitch, pitch is the main criterion. On the other hand, one can postulate that the mechanisms involved are gradually established since childhood through exposure to the environment, and thus the prosodic elements could take precedence in everyday life communication as it conveys information about the speaker's attitude (willingness to communicate, interest toward the interlocutors). Our study focuses on voice pleasantness and its relationship with social impression formation, exploring both the spectral aspects (pitch, timbre) and the prosodic ones. In our study, we recorded the voices through two vocal corpus (five vowels and a reading text) of 25 French males speaking French and 25 Iranian males speaking Farsi. French listeners (40 male/40 female) listened to the French voices and made a judgment either on the voice's pleasantness or on the speaker (judgment about his intelligence, honesty, sociability). The regression analyses from our acoustic measures showed that the prosodic elements (for example, the intonation and the speech rate) are the most important criteria concerning pleasantness, whatever the corpus or the listener's gender. Moreover, the correlation analyses showed that the speakers with the voices judged as the most pleasant are considered the most intelligent, sociable, and honest. The voices in Farsi have been judged by 80 other French listeners (40 male/40 female), and we found the same effect of intonation concerning the judgment of pleasantness with the corpus «vowel» whereas with the corpus «text» the pitch is more important than the prosody. It may suggest that voice perception contains some elements invariant across culture/language, whereas others are influenced by the cultural/linguistic background of the listener. Shortly in the future, Iranian people will be asked to listen either to the French voices for half of them or to the Farsi voices for the other half and produce the same judgments as the French listeners. This experimental design could potentially make it possible to distinguish what is linked to culture and what is linked to language in the case of differences in voice perception.

Keywords: cross-cultural psychology, impression formation, pleasantness, voice perception

Procedia PDF Downloads 73

Authors: Zunaira Zaman, David Bohannon

Abstract:

This study explores the application of machine learning (ML) for detecting security vulnerabilities in source code. The research aims to assist organizations with large application portfolios and limited security testing capabilities in prioritizing security activities. ML-based approaches offer benefits such as increased confidence scores, false positives and negatives tuning, and automated feedback. The initial approach using natural language processing techniques to extract features achieved 86% accuracy during the training phase but suffered from overfitting and performed poorly on unseen datasets during testing. To address these issues, the study proposes using the abstract syntax tree (AST) for Java and C++ codebases to capture code semantics and structure and generate path-context representations for each function. The Code2Vec model architecture is used to learn distributed representations of source code snippets for training a machine-learning classifier for vulnerability prediction. The study evaluates the performance of the proposed methodology using two datasets and compares the results with existing approaches. The Devign dataset yielded 60% accuracy in predicting vulnerable code snippets and helped resist overfitting, while the Juliet Test Suite predicted specific vulnerabilities such as OS-Command Injection, Cryptographic, and Cross-Site Scripting vulnerabilities. The Code2Vec model achieved 75% accuracy and a 98% recall rate in predicting OS-Command Injection vulnerabilities. The study concludes that even partial AST representations of source code can be useful for vulnerability prediction. The approach has the potential for automated intelligent analysis of source code, including vulnerability prediction on unseen source code. State-of-the-art models using natural language processing techniques and CNN models with ensemble modelling techniques did not generalize well on unseen data and faced overfitting issues. However, predicting vulnerabilities in source code using machine learning poses challenges such as high dimensionality and complexity of source code, imbalanced datasets, and identifying specific types of vulnerabilities. Future work will address these challenges and expand the scope of the research.

Keywords: code embeddings, neural networks, natural language processing, OS command injection, software security, code properties

Procedia PDF Downloads 112

Authors: Sahar B. Ahmed, M. Mostafa Said, Mona I. Mohamed

Abstract:

Streptozotocin (STZ) caused a significant increase in blood glucose and malondialdehyde (MDA) levels in serum accompanied by a significant decrease in blood reduced glutathione (GSH) and superoxide dismutase (SOD) activities. Also, ALT, AST, albumin and urea were markedly affected by STZ injection. The oral administration of Mango leaves extract (MLE) one hour before STZ injection was significantly improved the blood glucose level, ALT, AST activities, albumin and urea that associated with the regulation of MDA, GSH and SOD levels. The antimicrobial activity of MLE showed a significant inhibitory activity against multidrug resistant gram positive and gram negative bacteria isolated from patients in Egyptian hospitals especially Salmonella typhi and typhimurium. In conclusion, results revealed the antioxidant, hypoglycemic effect and antimicrobial potentials of MLE under investigation. Further studies will be needed to investigate the prolonged period of MLE administration and its possible side effects.

Keywords: aqueous extract of mango leaves, STZ, antioxidant, hypoglycemic effect, antimicrobial potentials.

Procedia PDF Downloads 462

Authors: Emilee M. Cruz

Abstract:

Teachers take part in short-term teaching fellowships abroad, and their preparation before, during, and after the experience is critical to affecting teachers’ feelings of success in the international classroom. A program evaluation of the teacher preparation within TALMA: The Israel Program for Excellence in English (TALMA) full-year teaching fellowship was conducted. A questionnaire was developed that examined professional development, deliberate reflection, and cultural and language immersion offered before, during, and after the short-term experience. The evaluation also surveyed teachers’ feelings of preparedness for the Israeli classroom and any recommendations they had for future teacher preparation within the fellowship program. The review suggests the TALMA program includes integrated professional learning communities between fellows and Israeli co-teachers, more opportunities for immersive Hebrew language learning, a broader professional network with Israelis, and opportunities for guided discussion with the TALMA community continued participation in TALMA events and learning following the full-year fellowship. Similar short-term international programs should consider the findings in the design of their participation preparation programs. The review also offers direction for future program evaluation of short-term participant preparation, including the need for frequent response item updates to match current offerings and evaluation of participant feelings of preparedness before, during, and after the full-year fellowship.

Keywords: educational program evaluation, international teaching, short-term teaching, teacher beliefs, teaching fellowship, teacher preparation

Procedia PDF Downloads 185

Authors: Sarina Zainab Shirazi

Abstract:

Abstract— With the advent of globalization, different countries have liberalized their trade policies to enhance economic integration and developmental processes but the advantages accrued vary greatly from region to region. This study specifically examines European Union (EU) and Association for South East Asian Nations (ASEAN), two regions that show contrasting integration patterns. EU shows most successful integrations versus the slower paced integration in the ASEAN region. A comprehensive panel data empirical investigation of EU and ASEAN in the context of economy size, geographical distances, language, ethnicity, common border and regional trade agreements (RTA) is conducted for a period of 1985 – 2015. The empirical investigation through the augmented gravity equation shows that the real effectiveness for enhanced intra-regional trade is significant when specific examination of export and import components is conducted in the presence of non-tariff barriers. These barriers surface in the form of terms of trade openness, inflation, exchange rate, common borders, common language, ethnic similarity, and presence of a formal regional trade agreement (RTA). Thus, these factors can be utilized by the EU and ASEAN regions in order to formulate effective policy tools to enhance trade within their respective spheres of influence.

Keywords: Association for South East Asian Nations, European Union, Gravity Model, Regional Trade

Procedia PDF Downloads 122

Authors: S. Nakayama

Abstract:

For non-native English speakers, using English modal auxiliary verbs appropriately can be among the most challenging tasks. This research sought to identify differences in modal verb usage between Japanese non-native English speakers (JNNS) and native speakers (NS) from two different perspectives: frequency of use and distribution of verb phrase structures (VPS) where modal verbs occur. This study can contribute to the identification of JNNSs' interlanguage with regard to modal verbs; the main aim is to make a suggestion for the improvement of teaching materials as well as to help language teachers to be able to teach modal verbs in a way that is helpful for learners. To address the primary question in this study, usage of nine central modals (‘can’, ‘could’, ‘may’, ‘might’, ‘shall’, ‘should’, ‘will’, ‘would’, and ‘must’) by JNNS was compared with that by NSs in the International Corpus Network of Asian Learners of English (ICNALE). This corpus is one of the largest freely-available corpora focusing on Asian English learners’ language use. The ICNALE corpus consists of four modules: ‘Spoken Monologue’, ‘Spoken Dialogue’, ‘Written Essays’, and ‘Edited Essays’. Among these, this research adopted the ‘Written Essays’ module only, which is the set of 200-300 word essays and contains approximately 1.3 million words in total. Frequency analysis revealed gaps as well as similarities in frequency order. Specifically, both JNNSs and NSs used ‘can’ with the most frequency, followed by ‘should’ and ‘will’; however, usage of all the other modals except for ‘shall’ was not identical to each other. A log-likelihood test uncovered JNNSs’ overuse of ‘can’ and ‘must’ as well as their underuse of ‘will’ and ‘would’. VPS analysis revealed that JNNSs used modal verbs in a relatively narrow range of VPSs as compared to NSs. Results showed that JNNSs used most of the modals with bare infinitives or the passive voice only whereas NSs used the modals in a wide range of VPSs including the progressive construction and the perfect aspect, both of which were the structures where JNNSs rarely used the modals. Results of frequency analysis suggest that language teachers or teaching materials should explain other modality items so that learners can avoid relying heavily on certain modals and have a wide range of lexical items to reflect their feelings more accurately. Besides, the underused modals should be more stressed in the classroom because they are members of epistemic modals, which allow us to not only interject our views into propositions but also build a relationship with readers. As for VPSs, teaching materials should present more examples of the modals occurring in a wide range of VPSs to help learners to be able to express their opinions from a variety of viewpoints.

Keywords: corpus linguistics, Japanese learners of English, modal auxiliary verbs, International Corpus Network of Asian Learners of English

Procedia PDF Downloads 129

Authors: Sabiha Sahin

Abstract:

Introduction: We have planned this study because of the increasing number of corrosive substance poisoning who admitted to the Pediatric Emergency Department. Method: 636 corrosive substance poisoning cases applied to the Osmangazi University PED between 1 January 2015 - 31 December 2017 were retrospectively analyzed. Results: 438 (69%) cases were male.The poisoning rate of groups 0-4 was significantly higher and occurred by accident. A total of 616 cases (96.9%) were poisoned by the oral route, and 20 cases were poisoned by inhalation.462 Patients were admitted to the hospital within an hour (72.8%). Of the 134 patients who had only erosion and redness around the mouth, 24 patients had salivation and dysphagia symptoms besides these. Of the 28 cases of 158 patients with symptoms, eusaphegeal stenosis was detected. on third day examination. Although there was no statistically significant correlation between esophageal stenosis and erosion and redness around the mouth, there was a statistically significant correlation between dysphagia and salivation between esophageal stenosis Conclusion: The increased salivation and the dysphagia are important signs of risk of devoloping esophageal stenosis at first examination corrosive poisoning in children.

Keywords: caustic, corrosive, poisoning, children

Procedia PDF Downloads 164

Authors: Aşina Gülerarslan, M. Faik Özdengül

Abstract:

The new economic, social and political world order has led to the emergence of a wide range of persuasion strategies and practices based on an ever expanding marketing axis that involves organizations, ideas and persons as well as products and services. It is seen that since the 1990's, a wide variety of competitive marketing ideas have been offered systematically to target audiences in the field of politics as in other fields. When the components of marketing are taken into consideration, all kinds of communication efforts involving “political leaders”, who are conceptualized as products in terms of political marketing, serve a process of social persuasion, which cannot be restricted to election periods only, and a manageable “image”. In this context, image, which is concerned with how the political product is perceived, involves not only the political discourses shared with the public but also all kinds of biographical information about the leader, the leader’s specific way of living and routines and his/her attitudes and behaviors in their private lives, and all these are regarded as components of the “product image”. While on the one hand the leader’s verbal or supra-verbal references serve the way the “spirit of the product” is perceived –just as in brand positioning- they also show their self-esteem levels, in other words how they perceive themselves on the other hand. Indeed, their self-esteem levels are evaluated in three fundamental categories in the “Functional Analysis”, namely parent, child and adult, and it is revealed that the words, tone of voice and body language a person uses makes it easy to understand at what self-esteem level that person is. In this context, words, tone of voice and body language, which provide important clues as to the “self” of the person, are also an indication of how political leaders evaluate both “themselves” and “the mass/audience” in the communication they establish with their audiences. When the matter is taken from the perspective of Turkey, the levels of self-esteem in the relationships that the political leaders establish with the masses are also important in revealing how our society is seen from the perspective of a specific leader. Since the leader is a part of the marketing strategy of a political party as a product, this evaluation is significant in terms of the forms of relationships between political institutions in our country with the society. In this study, the self-esteem level in the documentary entitled “Master’s Story”, where Recep Tayyip Erdoğan’s life history is told, is analyzed in the context of words, tone of voice and body language. Within the scope of the study, at what level of self-esteem Recep Tayyip Erdoğan was in the “Master’s Story”, a documentary broadcast on Beyaz TV, was investigated using the content analysis method. First, based on the Functional Analysis Literature, a transactional approach scale was created regarding parent, adult and child self-esteem levels. On the basis of this scale, the prime minister’s self-esteem level was determined in three basic groups, namely “tone of voice”, “the words he used” and “body language”. Descriptive analyses were made to the data within the framework of these criteria and at what self-esteem level the prime minister spoke throughout the documentary was revealed.

Keywords: political marketing, leader image, level of self-esteem, transactional approach

Procedia PDF Downloads 342

Authors: Michel Patiño-Sáenz, Martín Haissiner, Jorge Martínez-Cotrina, Daniel Pastor, Hernando Santamaría-García, Maria-Alejandra Tangarife, Agustin Ibáñez, Sandra Baez

Abstract:

It has been shown that gruesome descriptions of harm can increase the punishment given to a transgressor. This biasing effect is mediated by negative emotions, which are elicited upon the presentation of gruesome descriptions. However, there is a lack of studies inquiring the influence of such descriptions on moral decision-making in people involved in the criminal justice system. Such populations are of special interest since they have experience dealing with gruesome evidence, but also formal education on how to assess evidence and gauge the appropriate punishment according to the law. Likewise, they are expected to be objective and rational when performing their duty, because their decisions can impact profoundly people`s lives. Considering these antecedents, the objective of this study was to explore the influence gruesome written descriptions on moral decision-making in this group of people. To that end, we recruited attorneys, judges and public prosecutors (Criminal justice group, CJ, n=30) whose field of specialty is criminal law. In addition, we included a control group of people who did not have a formal education in law (n=30), but who were paired in age and years of education with the CJ group. All participants completed an online, Spanish-adapted version of a moral decision-making task, which was previously reported in the literature and also standardized and validated in the Latin-American context. A series of text-based stories describing two characters, one inflicting harm on the other, were presented to participants. Transgressor's intentionality (accidental vs. intentional harm) and language (gruesome vs. plain) used to describe harm were manipulated employing a within-subjects and a between-subjects design, respectively. After reading each story, participants were asked to rate (a) the harmful action's moral adequacy, (b) the amount of punishment deserving the transgressor and (c) how damaging was his behavior. Results showed main effects of group, intentionality and type of language on all dependent measures. In both groups, intentional harmful actions were rated as significantly less morally adequate, were punished more severely and were deemed as more damaging. Moreover, control subjects deemed more damaging and punished more severely any type of action than the CJ group. In addition, there was an interaction between intentionality and group. People in the control group rated harmful actions as less morally adequate than the CJ group, but only when the action was accidental. Also, there was an interaction between intentionality and language on punishment ratings. Controls punished more when harm was described using gruesome language. However, that was not the case of people in the CJ group, who assigned the same amount of punishment in both conditions. In conclusion, participants with job experience in the criminal justice system or criminal law differ in the way they make moral decisions. Particularly, it seems that they are less sensitive to the biasing effect of gruesome evidence, which is probably explained by their formal education or their experience in dealing with such evidence. Nonetheless, more studies are needed to determine the impact this phenomenon has on the fulfillment of their duty.

Keywords: criminal justice system, emotions, gruesome descriptions, intentionality, moral decision-making

Procedia PDF Downloads 194

Authors: Georgios Alexandropoulos

Abstract:

This study examines the rhetorical practice of "The consolation to himself upon the departure of the excellent Sallust" written by Flavius Claudius Julian the emperor. Its purpose is to describe the way that Julian uses the language as to have favorable effects on public through certain communicative and rhetorical functions.

Keywords: discourse analysis, Byzantine rhetoric,

Procedia PDF Downloads 418

Authors: Nina Dziatzko, Christopher Stehr, Franziska Struve

Abstract:

Intercultural competence as an explicit required skill nearly never appears in job advertisements in international or even global contexts. But especially those who have to deal with different nationalities and cultures in their everyday business need to have several intercultural competencies and further a global mind-set. This way the question arises how potential future global managers can be trained to learn these competencies. In this regard, it might be helpful to see if different types of intercultural trainings have different effects on those skills. This paper outlines lessons learned based on the evaluation of two different intercultural trainings for management students. The main differences between the observed intercultural trainings are the amount of theoretical input in relation to hands-on experiences, the number of trainers as well as the used methods to teach implicit cultural rules. Both groups contain management students with the willingness and perspective to work abroad or to work in international context. The research is carried out with a pre-training-survey and a post-training-survey which consists of questions referring the international context of the students and a self-estimation of 19 identified intercultural and global mind-set skills, such as: cosmopolitanism, empathy, differentiation and adaptability. Whereas there is no clear result which training gets overall a significant higher increase of skills, there is a clear difference between the focus of competencies trained by each of the intercultural trainings. This way this research provides a guideline for both academicals institutions as well as companies for the decision between different types of intercultural trainings, if the to be trained required skills are defined. Therefore the efficiency and the accuracy of fit of the education of future global managers get optimized.

Keywords: global mind-set, intercultural competencies, intercultural training, learning experiences

Procedia PDF Downloads 282

Authors: Raymond Xu, Ashley Hua, Andrew Wang, Yuru Lin

Abstract:

During COVID-19, the depression rate has increased dramatically. Young adults are most vulnerable to the mental health effects of the pandemic. Lower-income families have a higher ratio to be diagnosed with depression than the general population, but less access to clinics. This research aims to achieve early depression detection at low cost, large scale, and high accuracy with an interdisciplinary approach by incorporating clinical practices defined by American Psychiatric Association (APA) as well as multimodal AI framework. The proposed approach detected the nine depression symptoms with Natural Language Processing sentiment analysis and a symptom-based Lexicon uniquely designed for young adults. The experiments were conducted on the multimedia survey results from adolescents and young adults and unbiased Twitter communications. The result was further aggregated with the facial emotional cues analyzed by the Convolutional Neural Network on the multimedia survey videos. Five experiments each conducted on 10k data entries reached consistent results with an average accuracy of 88.31%, higher than the existing natural language analysis models. This approach can reach 300+ million daily active Twitter users and is highly accessible by low-income populations to promote early depression detection to raise awareness in adolescents and young adults and reveal complementary cues to assist clinical depression diagnosis.

Keywords: artificial intelligence, COVID-19, depression detection, psychiatric disorder

Procedia PDF Downloads 133

Authors: Wei Sun, Yan Dong

Abstract:

There is an increasing interest in introducing computational thinking at an early age. Computational thinking, like mathematical thinking, engineering thinking, and scientific thinking, is a kind of analytical thinking. Learning computational thinking skills is not only to improve technological literacy, but also allows learners to equip with practicable skills such as problem-solving skills. As people realize the importance of computational thinking, the field of educational technology faces a problem: how to choose appropriate tools and activities to help students develop computational thinking skills. Robots are gradually becoming a popular teaching tool, as robots provide a tangible way for young children to access to technology, and controlling a robot through programming offers them opportunities to engage in developing computational thinking. This study explores whether the introduction of flowcharts into the robotics programming courses can help children convert natural language into a programming language more easily, and then to better cultivate their computational thinking skills. An experimental study was adopted with a sample of children ages six to seven (N = 16) participated, and a one-meter-tall humanoid robot was used as the teaching tool. Results show that children can master basic programming concepts through robotic courses. Children's computational thinking has been significantly improved. Besides, results suggest that flowcharts do have an impact on young children’s computational thinking skills development, but it only has a significant effect on the "sequencing" and "correspondence" skills. Overall, the study demonstrates that the humanoid robot and flowcharts have qualities that foster young children to learn programming and develop computational thinking skills.

Keywords: robotics, computational thinking, programming, young children, flow chart

Procedia PDF Downloads 150

Authors: Jose Marie E. Ocdenaria, Riceli C. Mendoza

Abstract:

Culturally, languages meet, merge, share, exchange, appropriate, donate, and divide in and to and from each other. Further, this type of recurrence manifests in East Asian cultures, where language influence diffuses across geographical proximities. Historically, China has notable impacts on Japan’s culture. For instance, Japanese borrowed words from China and their way of reading and writing. This qualitative and descriptive employing contrastive analysis study addressed the true and false cognates of Japanese-Philippine languages and Chinese-Philippine languages. It involved a rich collection of data from various sources like textual pieces of evidence or corpora to gain a deeper understanding of true and false cognates between L1 and L2. Cognates of Japanese-Philippine languages and Chinese-Philippine languages were analyzed contrastively according to orthography, phonology, and semantics. The words presented were the roots; however, derivatives, reduplications, and variants of stress were included when they shed emphases on the comparison. The basis of grouping the cognates was its phonetic-semantic resemblance. Based on the analysis, it revealed that there are words which may have several types of lexical relationship. Further, the study revealed that the Japanese language has more false cognates in the Philippine languages, particularly in Tagalog and Cebuano. On the other hand, there are more true cognates of Chinese in Tagalog. It is the hope of this study to provide a significant contribution to a diverse audience. These include the teachers and learners of foreign languages such as Japanese and Chinese, future researchers and investigators, applied linguists, curricular theorists, community, and publishers.

Keywords: Contrastive Analysis, Japanese, Chinese and Philippine languages, Qualitative and descriptive study, True and False Cognates

Procedia PDF Downloads 141

Authors: Sheikh Shams Azam, Manoj Raju, Venkatesh Pagidimarri, Vamsi Kasivajjala

Abstract:

Over the past decade, there has been a steep rise in the data-driven analysis in major areas of medicine, such as clinical decision support system, survival analysis, patient similarity analysis, image analytics etc. Most of the data in the field are well-structured and available in numerical or categorical formats which can be used for experiments directly. But on the opposite end of the spectrum, there exists a wide expanse of data that is intractable for direct analysis owing to its unstructured nature which can be found in the form of discharge summaries, clinical notes, procedural notes which are in human written narrative format and neither have any relational model nor any standard grammatical structure. An important step in the utilization of these texts for such studies is to transform and process the data to retrieve structured information from the haystack of irrelevant data using information retrieval and data mining techniques. To address this problem, the authors present Q-Map in this paper, which is a simple yet robust system that can sift through massive datasets with unregulated formats to retrieve structured information aggressively and efficiently. It is backed by an effective mining technique which is based on a string matching algorithm that is indexed on curated knowledge sources, that is both fast and configurable. The authors also briefly examine its comparative performance with MetaMap, one of the most reputed tools for medical concepts retrieval and present the advantages the former displays over the latter.

Keywords: information retrieval, unified medical language system, syntax based analysis, natural language processing, medical informatics

Procedia PDF Downloads 137

Authors: Martina Katalin Szabó, Bernadett Lipóczi, Csenge Guba, István Uveges

Abstract:

In this study, an exhaustive analysis was carried out about the historical development of negative emotive intensifiers in the Hungarian language via NLP methods. Intensifiers are linguistic elements which modify or reinforce a variable character in the lexical unit they apply to. Therefore, intensifiers appear with other lexical items, such as adverbs, adjectives, verbs, infrequently with nouns. Due to the complexity of this phenomenon (set of sociolinguistic, semantic, and historical aspects), there are many lexical items which can operate as intensifiers. The group of intensifiers are admittedly one of the most rapidly changing elements in the language. From a linguistic point of view, particularly interesting are a special group of intensifiers, the so-called negative emotive intensifiers, that, on their own, without context, have semantic content that can be associated with negative emotion, but in particular cases, they may function as intensifiers (e.g.borzasztóanjó ’awfully good’, which means ’excellent’). Despite their special semantic features, negative emotive intensifiers are scarcely examined in literature based on large Historical corpora via NLP methods. In order to become better acquainted with trends over time concerning the intensifiers, The exhaustively analysed a specific historical corpus, namely the Magyar TörténetiSzövegtár (Hungarian Historical Corpus). This corpus (containing 3 millions text words) is a collection of texts of various genres and styles, produced between 1772 and 2010. Since the corpus consists of raw texts and does not contain any additional information about the language features of the data (such as stemming or morphological analysis), a large amount of manual work was required to process the data. Thus, based on a lexicon of negative emotive intensifiers compiled in a previous phase of the research, every occurrence of each intensifier was queried, and the results were stored in a separate data frame. Then, basic linguistic processing (POS-tagging, lemmatization etc.) was carried out automatically with the ‘magyarlanc’ NLP-toolkit. Finally, the frequency and collocation features of all the negative emotive words were automatically analyzed in the corpus. Outcomes of the research revealed in detail how these words have proceeded through grammaticalization over time, i.e., they change from lexical elements to grammatical ones, and they slowly go through a delexicalization process (their negative content diminishes over time). What is more, it was also pointed out which negative emotive intensifiers are at the same stage in this process in the same time period. Giving a closer look to the different domains of the analysed corpus, it also became certain that during this process, the pragmatic role’s importance increases: the newer use expresses the speaker's subjective, evaluative opinion at a certain level.

Keywords: historical corpus analysis, historical linguistics, negative emotive intensifiers, semantic changes over time

Procedia PDF Downloads 238

Authors: Mst Ilme Faridatul, Bo Wu

Abstract:

Multi-temporal urban land cover mapping is of paramount importance for monitoring urban sprawl and managing the ecological environment. For diversified urban activities, it is challenging to map land covers in a complex urban environment. Spectral indices have proved to be effective for mapping urban land covers. To improve multi-temporal urban land cover classification and mapping, we evaluate the performance of three spectral indices, e.g. modified normalized difference bare-land index (MNDBI), tasseled cap water and vegetation index (TCWVI) and shadow index (ShDI). The MNDBI is developed to evaluate its performance of enhancing urban impervious areas by separating bare lands. A tasseled cap index, TCWVI is developed to evaluate its competence to detect vegetation and water simultaneously. The ShDI is developed to maximize the spectral difference between shadows of skyscrapers and water and enhance water detection. First, this paper presents a comparative analysis of three spectral indices using Landsat Enhanced Thematic Mapper (ETM), Thematic Mapper (TM) and Operational Land Imager (OLI) data. Second, optimized thresholds of the spectral indices are imputed to classify land covers, and finally, their performance of enhancing multi-temporal urban land cover mapping is assessed. The results indicate that the spectral indices are competent to enhance multi-temporal urban land cover mapping and achieves an overall classification accuracy of 93-96%.

Keywords: land cover, mapping, multi-temporal, spectral indices

Procedia PDF Downloads 158

Authors: Adriane Orenha-Ottaiano

Abstract:

Collocations are conventionalized, recurrent and arbitrary lexical combinations. Due to the fact that they are highly specific for a particular language and may be contextually restricted, collocations pose a problem to EFL/ESL learners with regard to production or encoding. Taking that into account, the compilation of monolingual and bilingual collocations dictionaries for the referred audience is highly crucial and significant. Thus, the aim of this paper is to discuss the importance of the compilation of an Online Corpus-based Bilingual Collocations Dictionary, in the English-Portuguese and Portuguese-English directions. On a first phase, with the use of WordSmith Tools, the collocations were extracted from a Translation Learner Corpus (TLC), a parallel corpus made up of university students’ translations in the Portuguese-English direction, with approximately 100,000 words. In a second stage, based on the keywords analyzed from the TLC, more collocational patterns were extracted using the Sketch Engine. In order to include more collocations as well as to ensure dictionary users will have access to more frequent and recurrent collocations, we also use the frequency list from The Corpus of Contemporary American English, with the purpose of extracting more patterns. The dictionary focuses on all types of collocations (verbal, noun, adjectival and adverbial collocations), in order to help the referred audience use them more accurately and productively – so far the dictionary has more than 330 entries, and more than 3,500 collocations extracted. The idea of having the proposed dictionary in online format may allow to incorporate more qualitatively and quantitatively collocational information. Besides, more examples may be included, different from conventional printed collocations dictionaries. Being the first bilingual collocations dictionary in the aforementioned directions, it is hoped to achieve the challenge of meeting learners’ collocational needs as the collocations have been selected according to learners’ difficulties regarding the use of collocations.

Keywords: Corpus-Based Collocations Dictionary, Collocations , Bilingual Collocations Dictionary, Collocational Patterns

Procedia PDF Downloads 312

Authors: Inas Hussein

Abstract:

Communication in the modern world is increasingly becoming multimodal due to globalization and the digital space we live in which have remarkably affected how people communicate. Accordingly, Multimodal Discourse Analysis (MDA) is an emerging paradigm in discourse studies with the underlying assumption that other semiotic resources such as images, colours, scientific symbolism, gestures, actions, music and sound, etc. combine with language in order to  communicate meaning. One of the effective multimodal media that combines both verbal and non-verbal elements to create meaning is political cartoons. Furthermore, since political and social issues are mirrored in political cartoons, these are regarded as potential objects of discourse analysis since they not only reflect the thoughts of the public but they also have the power to influence them. The aim of this paper is to analyze some selected cartoons on the recognition of Jerusalem as Israel's capital by the American President, Donald Trump, adopting a multimodal approach. More specifically, the present research examines how the various semiotic tools and resources utilized by the cartoonists function in projecting the intended meaning. Ten political cartoons, among a surge of editorial cartoons highlighted by the Anti-Defamation League (ADL) - an international Jewish non-governmental organization based in the United States - as publications in different Arabic-language newspapers in Egypt, Saudi Arabia, UAE, Oman, Iran and UK, were purposively selected for semiotic analysis. These editorial cartoons, all published during 6^th–18^th December 2017, invariably suggest one theme: Jewish and Israeli domination of the United States. The data were analyzed using the framework of Visual Social Semiotics. In accordance with this methodological framework, the selected visual compositions were analyzed in terms of three aspects of meaning: representational, interactive and compositional. In analyzing the selected cartoons, an interpretative approach is being adopted. This approach prioritizes depth to breadth and enables insightful analyses of the chosen cartoons. The findings of the study reveal that semiotic resources are key elements of political cartoons due to the inherent political communication they convey. It is proved that adequate interpretation of the three aspects of meaning is a prerequisite for understanding the intended meaning of political cartoons. It is recommended that further research should be conducted to provide more insightful analyses of political cartoons from a multimodal perspective.

Keywords: Multimodal Discourse Analysis (MDA), multimodal text, political cartoons, visual modality

Procedia PDF Downloads 247

Authors: Pavel H. Llamocca, Victoria Lopez

Abstract:

The main objective of our work is the comparative analysis of environmental data from Open Data bases, belonging to different governments. This means that you have to integrate data from various different sources. Nowadays, many governments have the intention of publishing thousands of data sets for people and organizations to use them. In this way, the quantity of applications based on Open Data is increasing. However each government has its own procedures to publish its data, and it causes a variety of formats of data sets because there are no international standards to specify the formats of the data sets from Open Data bases. Due to this variety of formats, we must build a data integration process that is able to put together all kind of formats. There are some software tools developed in order to give support to the integration process, e.g. Data Tamer, Data Wrangler. The problem with these tools is that they need data scientist interaction to take part in the integration process as a final step. In our case we don’t want to depend on a data scientist, because environmental data are usually similar and these processes can be automated by programming. The main idea of our tool is to build Hadoop procedures adapted to data sources per each government in order to achieve an automated integration. Our work focus in environment data like temperature, energy consumption, air quality, solar radiation, speeds of wind, etc. Since 2 years, the government of Madrid is publishing its Open Data bases relative to environment indicators in real time. In the same way, other governments have published Open Data sets relative to the environment (like Andalucia or Bilbao). But all of those data sets have different formats and our solution is able to integrate all of them, furthermore it allows the user to make and visualize some analysis over the real-time data. Once the integration task is done, all the data from any government has the same format and the analysis process can be initiated in a computational better way. So the tool presented in this work has two goals: 1. Integration process; and 2. Graphic and analytic interface. As a first approach, the integration process was developed using Java and Oracle and the graphic and analytic interface with Java (jsp). However, in order to open our software tool, as second approach, we also developed an implementation with R language as mature open source technology. R is a really powerful open source programming language that allows us to process and analyze a huge amount of data with high performance. There are also some R libraries for the building of a graphic interface like shiny. A performance comparison between both implementations was made and no significant differences were found. In addition, our work provides with an Official Real-Time Integrated Data Set about Environment Data in Spain to any developer in order that they can build their own applications.

Keywords: open data, R language, data integration, environmental data

Procedia PDF Downloads 317

Authors: Mohd Adil, Rosina Khan, Asad U. Khan, Vasantha Rupasinghe HP

Abstract:

The study examined the effects of Syzygium aromaticum extracts on the virulence properties of Streptococcus mutans. The activity of glucosyltransferases in the presence of crude and diethylether fraction was reduced to 80% at concentration 78.12μg/ml and 39.06μg/ml respectively. The glycolytic pH drop by S. mutans cells was also disrupted by these extracts without affecting the bacterial viability. Microscopic analysis revealed morphological changes of the S. mutans biofilms, indicating that these plant extracts at sub-MICs could significantly affect the ability of S. mutans to form biofilm with distorted extracellular matrix. Furthermore, with the help of quantitative RT-PCR, the expression of different genes involved in adherence, quorum sensing, in the presence of these extracts were down regulated. The crude and active fractions were found effective in preventing caries development in rats. The data showed that S. aromaticum holds promise as a naturally occurring source of compounds that may prevent biofilm-related oral diseases.

Keywords: biofilm, quorum sensing, Streptococcus mutans, Syzygium aromaticum extract

Procedia PDF Downloads 312

Authors: Ragy Raafat Gaber Attaalla

Abstract:

Despite having the greatest rates of mortality and morbidity in the world, low- and middle-income (LMIC) nations trail high-income nations in terms of the number of clinical trials, the number of qualified researchers, and the amount of research information specific to their people. Health inequities and the use of precision medicine may be hampered by a lack of local genomic data, clinical pharmacology and pharmacometrics competence, and training opportunities. These issues can be solved by carrying out health care infrastructure development, which includes data gathering and well-designed clinical pharmacology training in LMICs. It will be advantageous if there is international cooperation focused at enhancing education and infrastructure and promoting locally motivated clinical trials and research. This paper outlines various instances where clinical pharmacology knowledge could be put to use, including pharmacogenomic opportunities that could lead to better clinical guideline recommendations. Examples of how clinical pharmacology training can be successfully implemented in LMICs are also provided, including clinical pharmacology and pharmacometrics training programmes in Africa and a Tanzanian researcher's personal experience while on a training sabbatical in the United States. These training initiatives will profit from advocacy for clinical pharmacologists' employment prospects and career development pathways, which are gradually becoming acknowledged and established in LMICs. The advancement of training and research infrastructure to increase clinical pharmacologists' knowledge in LMICs would be extremely beneficial because they have a significant role to play in global health.

Keywords: low- and middle-income, clinical pharmacology, pharmacometrics, career development pathways

Procedia PDF Downloads 75

Authors: Anna Wendolowska, Dorota Czyzowska

Abstract:

An intimate relationship is a significant factor that influences romantic partners’ well-being. In the face of stress, avoidant partners often employ a defense-against-intimacy strategy, leading to reduced relationship satisfaction, intimacy, interdependence, and longevity. Dyadic coping can buffer the negative effects of stress on relational satisfaction. Emotional competence mediates the relationship between insecure attachment and intimacy. In the current study, the link between attachment, different forms of dyadic coping, and various aspects of relationship satisfaction was examined. Both partners completed the attachment style questionnaire, the well matching couple questionnaire, and the dyadic coping inventory. The data was analyzed using the actor–partner interdependence model. The results highlighted a negative association between insecure-avoidant attachment style and intimacy. The actor effects of avoidant attachment on relational intimacy for women and for men were significant, whilst the partner effects for both spouses were not significant. The emotion-focused common dyadic coping moderated the relationship between avoidance of attachment and the partner's sense of intimacy. After controlling for the emotion-focused common dyadic coping, the actor effect of attachment on intimacy for men was slightly weaker, and the actor effect for women turned out to be insignificant. The emotion-focused common dyadic coping weakened the negative association between insecure attachment and relational intimacy. The impact of adult attachment and dyadic coping significantly contributes to subjective relational well-being.

Keywords: adult attachment, dyadic coping, relational intimacy, relationship satisfaction

Procedia PDF Downloads 166

Authors: B. S. Premalatha, Kausalya Sahani

Abstract:

Background: Glomus jugulare tumors arise within the jugular foramen and are commonly seen in females particularly on the left side. Surgical excision of the tumor may cause lower cranial nerve deficits. Cranial nerve involvement produces hoarseness of voice, slurred speech, and dysphagia along with other physical symptoms, thereby affecting the quality of life of individuals. Though oncological clearance is mainly emphasized on while treating these individuals, little importance is given to their communication, voice and swallowing problems, which play a crucial part in daily functioning. Objective: To examine the functions of voice, speech and swallowing outcomes of the subjects, following excision of glomus jugulare tumor. Methods: Two female subjects aged 56 and 62 years had come with a complaint of change in voice, inability to swallow and reduced clarity of speech following surgery for left glomus jugulare tumor were participants of the study. Their surgical information revealed multiple cranial nerve palsies involving the left facial, left superior and recurrent branches of the vagus nerve, left pharyngeal, left soft palate, left hypoglossal and vestibular nerves. Functional outcomes of voice, speech and swallowing were evaluated by perceptual and objective assessment procedures. Assessment included the examination of oral structures and functions, dysarthria by Frenchey dysarthria assessment, cranial nerve functions and swallowing functions. MDVP and Dr. Speech software were used to evaluate acoustic parameters of voice and quality of voice respectively. Results: The study revealed that both the subjects, subsequent to excision of glomus jugulare tumor, showed a varied picture of affected oral structure and functions, articulation, voice and swallowing functions. The cranial nerve assessment showed impairment of the vagus, hypoglossal, facial and glossopharyngeal nerves. Voice examination indicated vocal cord paralysis associated with breathy quality of voice, weak voluntary cough, reduced pitch and loudness range, and poor respiratory support. Perturbation parameters as jitter, shimmer were affected along with s/z ratio indicative of voice fold pathology. Reduced MPD(Maximum Phonation Duration) of vowels indicated that disturbed coordination between respiratory and laryngeal systems. Hypernasality was found to be a prominent feature which reduced speech intelligibility. Imprecise articulation was seen in both the subjects as the hypoglossal nerve was affected following surgery. Injury to vagus, hypoglossal, gloss pharyngeal and facial nerves disturbed the function of swallowing. All the phases of swallow were affected. Aspiration was observed before and during the swallow, confirming the oropharyngeal dysphagia. All the subsystems were affected as per Frenchey Dysarthria Assessment signifying the diagnosis of flaccid dysarthria. Conclusion: There is an observable communication and swallowing difficulty seen following excision of glomus jugulare tumor. Even with complete resection, extensive rehabilitation may be necessary due to significant lower cranial nerve dysfunction. The finding of the present study stresses the need for involvement of as speech and swallowing therapist for pre-operative counseling and assessment of functional outcomes.

Keywords: functional outcome, glomus jugulare tumor excision, multiple cranial nerve impairment, speech and swallowing

Procedia PDF Downloads 253

Authors: Sawitree Wongkittirungrueang

Abstract:

The objectives of this research were: 1) to study the effect of social stories on social interaction of students with autism. The sample was Pratomsuksa level 5 student with autism, Khon Kaen University Demonstration School, who was diagnosed by the Physician as High Functioning Autism since he was able to read, write, calculate and was studying in inclusive classroom. However, he still had disability in social interaction to participate in social activity group and communication. He could not learn how to develop friendship or create relationship. He had inappropriate behavior in social context. He did not understand complex social situations. In addition, he did seemed not know time and place. He was not able to understand feeling of oneself as well as the others. Consequently, he could not express his emotion appropriately. He did not understand or express his non-verbal language for communicating with friends. He lacked of common interest or emotion with nearby persons. He greeted inappropriately or was not interested in greeting. In addition, he did not have eye contact. He used inadequate language etc. He was elected by Purposive Sampling. His parents were willing to allow them to participate in this study. The research instruments were the lesson plan of social stories, and the picture book of social stories. The instruments used for data collection, were the social interaction evaluation of autistic students. This research was Quasi Experimental Research as One Group Pre-test, Post-test Design. For the Pre-test, the experiment was conducted by social stories. Then, the Post-test was implemented. The statistic used for data analysis, included the Mean, and Standard Deviation. The research findings were shown by Graph. The findings revealed hat the autistic students taught by social stories indicated better social interaction after being taught by social stories.

Keywords: social story, autism spectrum disorder (ASD), autism, social interaction

Procedia PDF Downloads 249

Authors: Ahmad Nazrun Shuid, Isa Naina Mohamed, Nurul Izzah Ibrahim, Norazlina Mohamed

Abstract:

Studies have shown that oral tocotrienol, a potent vitamin E, promoted fracture healing of osteoporotic bone. In this study, tocotrienol was combined with a polymer carrier (PLGA), and injected to the fracture site. The slow and constant release of tocotrienol particles would promote fracture healing of post-menopausal osteoporosis rat model. Fracture healing was assessed using micro-CT. Twenty-four Sprague-Dawley rats were ovariectomised or sham-operated and the left tibiae were fractured and fixed with plate and screws. The fractures were created at the upper third of the left tibiae. The rats were divided into 3 groups: sham-operated (SO), ovariectomised-control (OVxC) and PLGA-incorporated tocotrienol treatment (OVx + TT) groups. After 4 weeks, the OVx + TT group showed significantly better callus fracture healing than the OVxC group. In conclusion, tocotrienol-incorporated PLGA was able to promote fracture healing of osteoporotic bone.

Keywords: osteoporosis, micro-CT, tocotrienol, PLGA, fracture

Procedia PDF Downloads 670

Authors: Lisa Graham, Kathleen Grant

Abstract:

South Africa is a multicultural country with eleven official languages, yet most of the formal education at institutions of higher education in the country is in English. It is well known that many students, irrespective of their home language, struggle to grasp difficult scientific concepts and the same is true for students enrolled in the Extended Curriculum Programme at the Cape Peninsula University of Technology (CPUT), studying biomedical sciences. Today we are fortunate in that there is a plethora of resources available to students to research and better understand subject matter online. For example, the students often use YouTube videos to supplement the formal education provided in our course. Unfortunately, most of this material is presented in English. The rationale behind this project lies in that it is well documented that students think and grasp concepts easier in their home language and addresses the fact that the lingua franca of instruction in the field of biomedical science is English. A project aimed at addressing the lack of available resources in most of the South African languages is planned, where students studying Bachelor of Health Science in Medical Laboratory Science will collaborate with those studying Film and Video Technology to create educational videos, explaining scientific concepts in their home languages. These videos will then be published on our own YouTube channel, thereby making them accessible to fellow students, future students and anybody with interest in the subject. Research will be conducted to determine the benefit of the project as well as the published videos to the student community. It is suspected that the students engaged in making the videos will benefit in such a way as to gain further understanding of their course content, a broader appreciation of the discipline, an enhanced sense of civic responsibility, as well as greater respect for the different languages and cultures in our classes. Indeed, an increase in student engagement has been shown to play a central role in student success, and it is well noted that deeper learning and more innovative solutions take place in collaborative groups. We aim to make a meaningful contribution towards the production and repository of knowledge in multilingual teaching and learning for the benefit of the diverse student population and staff. This would strengthen language development, multilingualism, and multiculturalism at CPUT and empower and promote African languages as languages of science and education at CPUT, in other institutions of higher learning, and in South Africa as a whole.

Keywords: educational videos, multiculturalism, multilingualism, student engagement

Procedia PDF Downloads 159

Authors: Cherry Yieng Siang Ling, Joong Hee Lee, Myung Hwan Yun

Abstract:

The quality of a product to be usable has become the basic requirement in consumer’s perspective while failing the requirement ends up the customer from not using the product. Identifying usability issues from analyzing quantitative and qualitative data collected from usability testing and evaluation activities aids in the process of product design, yet the lack of studies and researches regarding analysis methodologies in qualitative text data of usability field inhibits the potential of these data for more useful applications. While the possibility of analyzing qualitative text data found with the rapid development of data analysis studies such as natural language processing field in understanding human language in computer, and machine learning field in providing predictive model and clustering tool. Therefore, this research aims to study the application capability of text processing algorithm in analysis of qualitative text data collected from usability activities. This research utilized datasets collected from LG neckband headset usability experiment in which the datasets consist of headset survey text data, subject’s data and product physical data. In the analysis procedure, which integrated with the text-processing algorithm, the process includes training of comments onto vector space, labeling them with the subject and product physical feature data, and clustering to validate the result of comment vector clustering. The result shows 'volume and music control button' as the usability feature that matches best with the cluster of comment vectors where centroid comments of a cluster emphasized more on button positions, while centroid comments of the other cluster emphasized more on button interface issues. When volume and music control buttons are designed separately, the participant experienced less confusion, and thus, the comments mentioned only about the buttons' positions. While in the situation where the volume and music control buttons are designed as a single button, the participants experienced interface issues regarding the buttons such as operating methods of functions and confusion of functions' buttons. The relevance of the cluster centroid comments with the extracted feature explained the capability of text processing algorithms in analyzing qualitative text data from usability testing and evaluations.

Keywords: usability, qualitative data, text-processing algorithm, natural language processing

Procedia PDF Downloads 286

Authors: B. R. Campomanes-Alvarez, P. Quiros, B. Fernandez

Abstract:

Automatic Speech Recognition (ASR) is a machine-based process of decoding and transcribing oral speech. A typical ASR system receives acoustic input from a speaker or an audio file, analyzes it using algorithms, and produces an output in the form of a text. Some speech recognition systems use Hidden Markov Models (HMMs) to deal with the temporal variability of speech and Gaussian Mixture Models (GMMs) to determine how well each state of each HMM fits a short window of frames of coefficients that represents the acoustic input. Another way to evaluate the fit is to use a feed-forward neural network that takes several frames of coefficients as input and produces posterior probabilities over HMM states as output. Deep neural networks (DNNs) that have many hidden layers and are trained using new methods have been shown to outperform GMMs on a variety of speech recognition systems. Acoustic models for state-of-the-art ASR systems are usually training on massive amounts of data. However, audio files with their corresponding transcriptions can be difficult to obtain, especially in the Spanish language. Hence, in the case of these low-resource scenarios, building an ASR model is considered as a complex task due to the lack of labeled data, resulting in an under-trained system. Semi-supervised learning approaches arise as necessary tasks given the high cost of transcribing audio data. The main goal of this proposal is to develop a procedure based on acoustic semi-supervised learning for Spanish ASR systems by using DNNs. This semi-supervised learning approach consists of: (a) Training a seed ASR model with a DNN using a set of audios and their respective transcriptions. A DNN with a one-hidden-layer network was initialized; increasing the number of hidden layers in training, to a five. A refinement, which consisted of the weight matrix plus bias term and a Stochastic Gradient Descent (SGD) training were also performed. The objective function was the cross-entropy criterion. (b) Decoding/testing a set of unlabeled data with the obtained seed model. (c) Selecting a suitable subset of the validated data to retrain the seed model, thereby improving its performance on the target test set. To choose the most precise transcriptions, three confidence scores or metrics, regarding the lattice concept (based on the graph cost, the acoustic cost and a combination of both), was performed as selection technique. The performance of the ASR system will be calculated by means of the Word Error Rate (WER). The test dataset was renewed in order to extract the new transcriptions added to the training dataset. Some experiments were carried out in order to select the best ASR results. A comparison between a GMM-based model without retraining and the DNN proposed system was also made under the same conditions. Results showed that the semi-supervised ASR-model based on DNNs outperformed the GMM-model, in terms of WER, in all tested cases. The best result obtained an improvement of 6% relative WER. Hence, these promising results suggest that the proposed technique could be suitable for building ASR models in low-resource environments.

Keywords: automatic speech recognition, deep neural networks, machine learning, semi-supervised learning

Procedia PDF Downloads 344