Search results for: pre-trained language models

Authors: Louisa Han Lin Tan, Poh Sim Kang, Wei Ming Loi, Susan Jane Rickard Liow

Abstract:

The efficacy of music for improving speech and language has been shown across ages and diagnoses. Across the world, the wide range of therapy settings and increasing number of children diagnosed with special needs demand more cost and time effective service delivery. However, research exploring co-treatment models on children other than those with Autism Spectrum Disorder remains sparse. The aim of this research was to determine the efficacy of music for improving language in children with special needs, and generalizability of therapy effects. 25 children (7 to 12 years) were split into three groups – A, B and control. A cross-over design with direct therapy (storytelling) with or without music, and indirect therapy was applied with two therapy phases lasting 6 sessions each. Therapy targeted three prepositions in each phase. Baseline language abilities were assessed, with re-assessment after each phase. The introduction of music in therapy led to significantly greater improvement (p=.046, r=.53) in associated language abilities, with case studies showing greater effectiveness in developmentally appropriate target prepositions. However, improvements were not maintained once direct therapy ceased. As such, the incorporation of music could lead to greater efficiency and effectiveness of language therapy in children with special needs, but sustainability and generalizability of therapy effects both require further exploration.

Keywords: music, language therapy, children, special needs

Procedia PDF Downloads 440

Authors: Yanqi Yang, Chongshan Liao, Cheuk Hin Ho, Susan Bridges

Abstract:

Aim: To investigate student’s perceptions of using e-models in an inquiry-based curriculum. Approach: 52 second-year dental students completed a pre- and post-test questionnaire relating to their perceptions of e-models and their use in inquiry-based learning. The pre-test occurred prior to any learning with e-models. The follow-up survey was conducted after one year's experience of using e-models. Results: There was no significant difference between the two sets of questionnaires regarding student’s perceptions of the usefulness of e-models and their willingness to use e-models in future inquiry-based learning. Most of the students preferred using both plaster models and e-models in tandem. Conclusion: Students did not change their attitude towards e-models and most of them agreed or were neutral that e-models are useful in inquiry-based learning. Whilst recognizing the utility of 3D models for learning, student's preference for combining these with solid models has implications for the development of haptic sensibility in an operative discipline.

Keywords: e-models, inquiry-based curriculum, education, questionnaire

Procedia PDF Downloads 401

Authors: Ozlem Bozok, Yusuf Bozok

Abstract:

The importance of using mother tongue and translation in foreign language classrooms cannot be ignored and translation can be utilized as a method in English Language Teaching courses. There exist researches advocating or objecting to the use of translation in foreign language learning but they all have a point in common: Translation should be used as an aid to teaching, not an end in itself. In this research, prospective English language teachers’ opinions about translation use and use of mother tongue in foreign language teaching are investigated and according to the findings, some explanations and recommendations are made.

Keywords: exposure to foreign language translation, foreign language learning, prospective teachers’ opinions, use of L1

Procedia PDF Downloads 502

Authors: Isaac K. E. Ampomah, Seong-Bae Park, Sang-Jo Lee

Abstract:

Over the past decade, there have been promising developments in Natural Language Processing (NLP) with several investigations of approaches focusing on Recognizing Textual Entailment (RTE). These models include models based on lexical similarities, models based on formal reasoning, and most recently deep neural models. In this paper, we present a sentence encoding model that exploits the sentence-to-sentence relation information for RTE. In terms of sentence modeling, Convolutional neural network (CNN) and recurrent neural networks (RNNs) adopt different approaches. RNNs are known to be well suited for sequence modeling, whilst CNN is suited for the extraction of n-gram features through the filters and can learn ranges of relations via the pooling mechanism. We combine the strength of RNN and CNN as stated above to present a unified model for the RTE task. Our model basically combines relation vectors computed from the phrasal representation of each sentence and final encoded sentence representations. Firstly, we pass each sentence through a convolutional layer to extract a sequence of higher-level phrase representation for each sentence from which the first relation vector is computed. Secondly, the phrasal representation of each sentence from the convolutional layer is fed into a Bidirectional Long Short Term Memory (Bi-LSTM) to obtain the final sentence representations from which a second relation vector is computed. The relations vectors are combined and then used in then used in the same fashion as attention mechanism over the Bi-LSTM outputs to yield the final sentence representations for the classification. Experiment on the Stanford Natural Language Inference (SNLI) corpus suggests that this is a promising technique for RTE.

Keywords: deep neural models, natural language inference, recognizing textual entailment (RTE), sentence-to-sentence relation

Procedia PDF Downloads 325

Authors: Rim Bouhaouel, Naoufel Kraïem, Zuhoor Al Khanjari

Abstract:

Unified Modeling Language (UML) is considered as one of the widespread modeling language standardized by the Object Management Group (OMG). Therefore, the model driving engineering (MDE) community attempts to provide reuse of UML diagrams, and do not construct it from scratch. The UML model appears according to a specific software development process. The existing method generation models focused on the different techniques of transformation without considering the development process. Our work aims to construct an UML component from fragments of UML diagram basing on an agile method. We define UML fragment as a portion of a UML diagram, which express a business target. To guide the generation of fragments of UML models using an agile process, we need a flexible approach, which adapts to the agile changes and covers all its activities. We use the software product line (SPL) to derive a fragment of process agile method. This paper explains our approach, named RECUP, to generate UML fragments following an agile process, and overviews the different aspects. In this paper, we present the approach and we define the different phases and artifacts.

Keywords: UML, component, fragment, agile, SPL

Procedia PDF Downloads 368

Authors: Wajiha Fatima

Abstract:

Various approaches and methods for second language classroom teaching have been proposed since the nineteenth century. Task Based Language Teaching has been prevailing approach in a second language classroom context. It is an approach which immerses students in a naturalistic setting. Tasks are the core unit of planning and instruction. This paper aims at expounding the concept of Task Based Language Teaching and how it has been evolved. In this study, researcher will highlight the usefulness of TBLT and the role it played as a powerful tool for learning and teaching in a second language setting. The article will reflect the implementation of various tasks based activities as well as the roles played by learners and teachers and the problems faced by them. In the end, researcher will discuss how TBLT can be implemented in second language classroom pedagogy.

Keywords: implementation, second language classroom, tasks, task based language teaching

Procedia PDF Downloads 323

Authors: Othmane Touri, Sanaa El Filali, El Habib Benlahmar

Abstract:

Natural Language Processing (NLP) has gained significant attention lately. It has proved its ability to analyze and extract insights from unstructured text data in various languages. It is found that one of the most popular NLP applications is sentiment analysis which aims to identify the sentiment expressed in a piece of text, such as positive, negative, or neutral, in multiple languages. While there are several multilingual NLP models available for sentiment analysis, there is a need to investigate their effectiveness in different contexts and applications. In this study, we aim to investigate the effectiveness of different multilingual NLP models for sentiment analysis on a dataset of online product reviews in multiple languages. The performance of several NLP models, including Google Cloud Natural Language API, Microsoft Azure Cognitive Services, Amazon Comprehend, Stanford CoreNLP, spaCy, and Hugging Face Transformers are being compared. The models based on several metrics, including accuracy, precision, recall, and F1 score, are being evaluated and compared to their performance across different categories of product reviews. In order to run the study, preprocessing of the dataset has been performed by cleaning and tokenizing the text data in multiple languages. Then training and testing each model has been applied using a cross-validation approach where randomly dividing the dataset into training and testing sets and repeating the process multiple times has been used. A grid search approach to optimize the hyperparameters of each model and select the best-performing model for each category of product reviews and language has been applied. The findings of this study provide insights into the effectiveness of different multilingual NLP models for Multilingual Sentiment Analysis and their suitability for different languages and applications. The strengths and limitations of each model were identified, and recommendations for selecting the most performant model based on the specific requirements of a project were provided. This study contributes to the advancement of research methods in multilingual NLP and provides a practical guide for researchers and practitioners in the field.

Keywords: NLP, multilingual, sentiment analysis, texts

Procedia PDF Downloads 61

Authors: Ammarah Irum, Muhammad Ali Tahir

Abstract:

Document-level sentiment analysis in Urdu is a challenging Natural Language Processing (NLP) task due to the difficulty of working with lengthy texts in a language with constrained resources. Deep learning models, which are complex neural network architectures, are well-suited to text-based applications in addition to data formats like audio, image, and video. To investigate the potential of deep learning for Urdu sentiment analysis, we implemented five different deep learning models, including Bidirectional Long Short Term Memory (BiLSTM), Convolutional Neural Network (CNN), Convolutional Neural Network with Bidirectional Long Short Term Memory (CNN-BiLSTM), and Bidirectional Encoder Representation from Transformer (BERT). In this study, we developed a hybrid deep learning model called BiLSTM-Single Layer Multi Filter Convolutional Neural Network (BiLSTM-SLMFCNN) by fusing BiLSTM and CNN architecture. The proposed and baseline techniques are applied on Urdu Customer Support data set and IMDB Urdu movie review data set by using pre-trained Urdu word embedding that are suitable for sentiment analysis at the document level. Results of these techniques are evaluated and our proposed model outperforms all other deep learning techniques for Urdu sentiment analysis. BiLSTM-SLMFCNN outperformed the baseline deep learning models and achieved 83%, 79%, 83% and 94% accuracy on small, medium and large sized IMDB Urdu movie review data set and Urdu Customer Support data set respectively.

Keywords: urdu sentiment analysis, deep learning, natural language processing, opinion mining, low-resource language

Procedia PDF Downloads 44

Authors: Adeep Hande, Shubham Agarwal

Abstract:

This paper presents a study on identifying sexism in online texts using various state-of-the-art deep learning models based on BERT. We experimented with different feature sets and model architectures and evaluated their performance using precision, recall, F1 score, and accuracy metrics. We also explored the use of pseudolabeling technique to improve model performance. Our experiments show that the best-performing models were based on BERT, and their multilingual model achieved an F1 score of 0.83. Furthermore, the use of pseudolabeling significantly improved the performance of the BERT-based models, with the best results achieved using the pseudolabeling technique. Our findings suggest that BERT-based models with pseudolabeling hold great promise for identifying sexism in online texts with high accuracy.

Keywords: large language models, semi-supervised learning, sexism detection, data sparsity

Procedia PDF Downloads 45

Authors: Lee Bih Ni, Kiu Su Na

Abstract:

This paper discusses the Chinese Language Teaching as a Second Language by focusing on Immersion Teaching. Researchers used narrative literature review to describe the current states of both art and science in focused areas of inquiry. Immersion teaching comes with a standard that teachers must reliably meet. Chinese language-immersion instruction consists of language and content lessons, including functional usage of the language, academic language, authentic language, and correct Chinese sociocultural language. Researchers used narrative literature reviews to build a scientific knowledge base. Researchers collected all the important points of discussion, and put them here with reference to the specific field where this paper is originally based on. The findings show that Chinese Language in immersion teaching is not like standard foreign language classroom; immersion setting provides more opportunities to teach students colloquial language than academic. Immersion techniques also introduce a language’s cultural and social contexts in a meaningful and memorable way. It is particularly important that immersion teachers connect classwork with real-life experiences. Immersion also includes more elements of discovery and inquiry based learning than do other kinds of instructional practices. Students are always and consistently interpreted the conclusions and context clues.

Keywords: a second language, Chinese language teaching, immersion teaching, instructional strategies

Procedia PDF Downloads 428

Authors: Vanessa Matias Leite, Jandira Guenka Palma, Flávio Henrique de Oliveira

Abstract:

This paper presents a model-oriented development approach to software development in the Model-View-Controller (MVC) architectural standard. This approach aims to expose a process of extractions of information from the models, in which through rules and syntax defined in this work, assists in the design of the initial model and its future conversions. The proposed paper presents a syntax based on the natural language, according to the rules agreed in the classic grammar of the Portuguese language, added to the rules of conversions generating models that follow the norms of the Object Management Group (OMG) and the Meta-Object Facility MOF.

Keywords: BNF Syntax, model driven architecture, model-view-controller, transformation, UML

Procedia PDF Downloads 369

Authors: Zhaoxin Luo, Michael Zhu

Abstract:

In natural languages, there are always complex semantic hierarchies. Obtaining the feature representation based on these complex semantic hierarchies becomes the key to the success of the model. Several RNN models have recently been proposed to use latent indicators to obtain the hierarchical structure of documents. However, the model that only uses a single-layer latent indicator cannot achieve the true hierarchical structure of the language, especially a complex language like Chinese. In this paper, we propose a deep layered model that stacks arbitrarily many RNN layers equipped with latent indicators. After using EM and training it hierarchically, our model solves the computational problem of stacking RNN layers and makes it possible to stack arbitrarily many RNN layers. Our deep hierarchical model not only achieves comparable results to large pre-trained models on the Chinese short text classification problem but also achieves state of art results on the Chinese long text classification problem.

Keywords: nature language processing, recurrent neural network, hierarchical structure, document classification, Chinese

Procedia PDF Downloads 41

Authors: Numra Qayyum, Samina Sarwat, Noor ul Ain

Abstract:

Bilingual speakers generally can speak both languages with the same competency without mixing them intentionally and making mistakes, but sometimes errors occur in language selection. This quantitative study particularly deals with the language errors made by Urdu-English bilinguals. In this research, researchers have given special attention to the part played by bottom-up priming and top-down cognitive control in these errors. Unstable Urdu-English bilingual participants termed pictures and were prompted to shift from one language to another under the pressure of time. Different situations were given to manipulate the participants. The long and short runs trials of the same language were also given before switching to another language. The study is concluded with the findings that bilinguals made more errors when switching to the first language from their second language, and these errors are large in number, especially when a speaker is switching from L2 (second language) to L1 (first language) after a long run. When the switching is reversed, i.e., from L2 to LI, it had no effect at all. These results gave the clear responsibility of all these errors to top-down cognitive control.

Keywords: bottom up priming, language error, language switching, top down cognitive control

Procedia PDF Downloads 111

Authors: Aaidha Hammad

Abstract:

The age a child to be exposed to a second language is a controversial issue in communities such as the Maldives where English is taught as a second language. It has been observed that different stakeholders have different viewpoints towards the issue. Some believe that the earlier children are exposed to a second language, the better they learn, while others disagree with the notion. Hence, this case study investigates whether children learn a second language better when they are exposed at an earlier age or not. The spoken and written data collected confirm that earlier exposure helps in mastering the sound pattern and speaking fluency with more native-like accent, while a later age is better for learning more abstract and concrete aspects such as grammar and syntactic rules.

Keywords: age, fluency, second language acquisition, development of language skills

Procedia PDF Downloads 386

Authors: Dinda N. Lestari

Abstract:

Sundanese language is one of heritage language in Indonesia that must be maintained especially the formal type of it because teenagers nowadays do not speak Sundanese language formally in their daily lives. To maintain it, Cultural and Education Ministry of Indonesia has input Sundanese language lesson at senior high school in West Java area. The aim of this study was to observe whether the existence of Sundanese language lesson in senior high school in the big town of Karawang, West Java - Indonesia give the contribution to the formal type of Sundanese language maintenance or not. For gathering the data, the researcher interviewed the senior high school students who have learned Sundanese language to observe their acquisition of it. As a result of the interview, the data was presented in qualitative research by using the interviewing method. Then, the finding indicated that the existence of Sundanese language in Senior High School also the educational program which is related to it, for instance, Kemis Nyunda seemed to do not effective enough in maintaining the formal type of Sundanese language. Therefore, West Java government must revise the learning strategy of it, including the role of the Sundanese language teacher.

Keywords: heritage language, language maintenance and shift, senior high school, Sundanese language, Sundanese language lesson

Procedia PDF Downloads 127

Authors: Edith Lotachukwu Ochege

Abstract:

Nations stay together when citizens share enough values and preferences and can communicate with each other. Homogeneity among people can be built with education, teaching a common language to facilitate communication, infrastructure for easier travel, but also by brute force such as prohibiting local cultures. This paper discusses the role of language education in nation building. It defines education, highlights the functions of language. Furthermore, it expresses socialization agents that aid culture which are all embodied in language, problems of nation building.

Keywords: nation building, language education, function of language, socialization

Procedia PDF Downloads 544

Authors: Oleg Redkin, Olga Bernikova

Abstract:

Globalization and migrations have made cross-cultural contacts more frequent and intensive. Sometimes, these contacts may lead to misunderstanding between partners of communication and misinterpretations of the verbal messages that some researchers tend to consider as the 'clash of civilizations'. In most cases, reasons for that may be found in cultural and linguistic differences and hence misinterpretations of intentions and behavior. The current research examines factors of verbal and non-verbal communication that should be taken into consideration in verbal and non-verbal contacts. Language is one of the most important manifestations of the cultural code, and it is often considered as one of the special features of a civilization. The Arabic language, in particular, is commonly associated with Islam and the language and the Arab-Muslim civilization. It is one of the most important markers of self-identification for more than 200 million of native speakers. Arabic is the language of the Quran and hence the symbol of religious affiliation for more than one billion Muslims around the globe. Adequate interpretation of Arabic texts requires profound knowledge of its grammar, semantics of its vocabulary. Communicating sides who belong to different cultural groups are guided by different models of behavior and hierarchy of values, besides that the vocabulary each of them uses in the dialogue may convey different semantic realities and vary in connotations. In this context direct, literal translation in most cases cannot adequately convey the original meaning of the original message. Besides that peculiarities and diversities of the extralinguistic information, such as the body language, communicative etiquette, cultural background and religious affiliations may make the dialogue even more difficult. It is very likely that the so called 'clash of civilizations' in most cases is due to misinterpretation of counterpart's means of discourse such as language, cultural codes, and models of behavior rather than lies in basic contradictions between partners of communication. In the process of communication, one has to rely on universal values rather than focus on cultural or religious peculiarities, to take into account current linguistic and extralinguistic context.

Keywords: Arabic, civilization, discourse, language, linguistic

Procedia PDF Downloads 197

Authors: Ratna Elizabeth

Abstract:

Motivation is a very important factor since it contributes a lot to the students’ success in learning a language. Using authentic materials is believed as a mean of increasing the motivation. The materials define as authentic if they are not specifically written for the purpose of language teaching. They are genuine spoken or written language data which are drawn from many different sources. The intention of this study is to investigate the impact of using of authentic materials on students’ motivation. A single case study is conducted to the grade 9 students who learn Indonesian Language as a Foreign Language (ILFL) at an international school in Jakarta, Indonesia. Questionnaires are also distributed to the students to know their perceptions on the using of authentic materials. The results show that the using of authentic materials has increased the students’ motivation in learning the language.

Keywords: authentic materials, ILFL, language learning, motivation

Procedia PDF Downloads 357

Authors: Bunakiye R. Japheth, Ogude U. Cyril

Abstract:

Since the process of transforming user requirements to modeling constructs are not very well supported by domain-specific frameworks, it became necessary to integrate domain requirements with the specific architectures to achieve an integrated customizable solutions space via artifact orientation. Domain-specific modeling language specifications of model-driven engineering technologies focus more on requirements within a particular domain, which can be tailored to aid the domain expert in expressing domain concepts effectively. Modeling processes through domain-specific language formalisms are highly volatile due to dependencies on domain concepts or used process models. A capable solution is given by artifact orientation that stresses on the results rather than expressing a strict dependence on complicated platforms for model creation and development. Based on this premise, domain-specific methods for producing artifacts without having to take into account the complexity and variability of platforms for model definitions can be integrated to support customizable development. In this paper, we discuss methods for the integration capabilities and necessities within a common structure and semantics that contribute a metamodel for artifact-orientation, which leads to a reusable software layer with concrete syntax capable of determining design intents from domain expert. These concepts forming the language formalism are established from models explained within the oil and gas pipelines industry.

Keywords: control process, metrics of engineering, structured abstraction, semantic model

Procedia PDF Downloads 117

Authors: Kai Zhu, Shanhua He, Yujiao Chang

Abstract:

This paper explores the virtual linguistic landscape of an emerging virtual language community in TikTok, a language community realizing immediate and non-immediate communication without a precise Spatio-temporal domain or a specific socio-cultural boundary or interpersonal network. This kind of language community generates a large number and various forms of virtual linguistic landscape, with which we conducted a virtual ethnographic survey together with telephone interviews to collect data from coping. We have been following two language communities in TikTok for several months so that we can illustrate the composition of the two language communities and some typical virtual language landscapes in both language communities first. Then we try to explore the reasons why and how they are formed through the organization, transcription, and analysis of the interviews. Our analysis reveals the richness and diversity of the virtual linguistic landscape, and finally, we summarize some of the characteristics of this language community.

Keywords: virtual linguistic landscape, virtual language community, virtual ethnographic survey, TikTok

Procedia PDF Downloads 79

Authors: Hsin Lee, Hsuan Lee

Abstract:

Legal artificial intelligence (LegalAI) has been increasing applications within legal systems, propelled by advancements in natural language processing (NLP). Compared with general documents, legal case documents are typically long text sequences with intrinsic logical structures. Most existing language models have difficulty understanding the long-distance dependencies between different structures. Another unique challenge is that while the Judiciary of Taiwan has released legal judgments from various levels of courts over the years, there remains a significant obstacle in the lack of labeled datasets. This deficiency makes it difficult to train models with strong generalization capabilities, as well as accurately evaluate model performance. To date, models in Taiwan have yet to be specifically trained on judgment data. Given these challenges, this research proposes a Longformer-based pre-trained language model explicitly devised for retrieving similar judgments in Taiwanese legal documents. This model is trained on a self-constructed dataset, which this research has independently labeled to measure judgment similarities, thereby addressing a void left by the lack of an existing labeled dataset for Taiwanese judgments. This research adopts strategies such as early stopping and gradient clipping to prevent overfitting and manage gradient explosion, respectively, thereby enhancing the model's performance. The model in this research is evaluated using both the dataset and the Average Entropy of Offense-charged Clustering (AEOC) metric, which utilizes the notion of similar case scenarios within the same type of legal cases. Our experimental results illustrate our model's significant advancements in handling similarity comparisons within extensive legal judgments. By enabling more efficient retrieval and analysis of legal case documents, our model holds the potential to facilitate legal research, aid legal decision-making, and contribute to the further development of LegalAI in Taiwan.

Keywords: legal artificial intelligence, computation and language, language model, Taiwanese legal cases

Procedia PDF Downloads 47

Authors: Promise Ekpo Osaine, Daniel Melesse

Abstract:

Image captioning is an active area of research in the multi-modal artificial intelligence (AI) community as it connects vision and language understanding, especially in settings where it is required that a model understands the content shown in an image and generates semantically and grammatically correct descriptions. In this project, we followed a standard approach to a deep learning-based image captioning model, injecting architecture for the encoder-decoder setup, where the encoder extracts image features, and the decoder generates a sequence of words that represents the image content. As such, we investigated image encoders, which are ResNet101, InceptionResNetV2, EfficientNetB7, EfficientNetV2M, and CLIP. As a caption generation structure, we explored long short-term memory (LSTM). The CLIP-LSTM model demonstrated superior performance compared to the encoder-decoder models, achieving a BLEU-1 score of 0.904 and a BLEU-4 score of 0.640. Additionally, among the CNN-LSTM models, EfficientNetV2M-LSTM exhibited the highest performance with a BLEU-1 score of 0.896 and a BLEU-4 score of 0.586 while using a single-layer LSTM.

Keywords: multi-modal AI systems, image captioning, encoder, decoder, BLUE score

Procedia PDF Downloads 35

Authors: Gareth Baxter

Abstract:

We use a mathematical model of language change to examine two recently observed patterns of language change: one in which most speakers change gradually, following the mean of the community change, and one in which most individuals use predominantly one variant or another, and change rapidly if they change at all. The model is based on Croft’s Utterance Selection account of language change, which views language change as an evolutionary process, in which different variants (different ‘ways of saying the same thing’) compete for usage in a population of speakers. Language change occurs when a new variant replaces an older one as the convention within a given population. The present model extends a previous simpler model to include effects related to speaker aging and interspeaker variation in behaviour. The two patterns of individual change (one more centralized and the other more polarized) were recently observed in historical language changes, and it was further observed that slower changes were more associated with the centralized pattern, while quicker changes were more polarized. Our model suggests that the two patterns of change can be explained by different balances between the preference of speakers to use one variant over another and the degree of accommodation to (propensity to adapt towards) other speakers. The correlation with the rate of change appears naturally in our model, and results from the fact that both differential weighting of variants and the degree of accommodation affect the time for change to occur, while also determining the patterns of change. This work represents part of an ongoing effort to examine phenomena in language change through the use of mathematical models. This offers another way to evaluate qualitative explanations that cannot be practically tested (or cannot be tested at all) in a real-world, large-scale speech community.

Keywords: agent based modeling, cultural evolution, language change, social behavior modeling, social influence

Procedia PDF Downloads 211

Authors: Touila Ahmed, Elie Louis, Hamza Gharbi

Abstract:

State of the art question answering models have recently shown impressive performance especially in a zero-shot setting. This approach is particularly useful when confronted with a highly diverse domain such as the legal field, in which it is increasingly difficult to have a dataset covering every notion and concept. In this work, we propose a flexible generative question answering approach to contract analysis as well as a weakly supervised procedure to leverage unannotated data and boost our models’ performance in general, and their zero-shot performance in particular.

Keywords: question answering, contract analysis, zero-shot, natural language processing, generative models, self-supervision

Procedia PDF Downloads 150

Authors: Mahrukh Baig

Abstract:

This research paper, sets out to draw some traditional english language teaching practices and to suggest ways for their improvement under the light of humanistic and socio-cultural psychology. This is going to aid language teachers by applying principled psychological methods on the field of education in order to introduce a reciprocal mode of teaching where teacher and learner begin with a mutual effort. However the teacher, after initiating most of the work, gradually passes on more and more responsibility to the learners resulting in their independent endeavors.

Keywords: English Language Teaching (ELT), Second Language Acquisition (SLA), teaching english as second/foreign language, humanistic psychology, socio-cultural psychology, application of psychology to language teaching

Procedia PDF Downloads 574

Authors: Bowen Ding, Yihao Kuang

Abstract:

In recent years, with the development of large-scale pre-trained lan-guage models, building vector representations of text through deep neural network models has become a standard practice for natural language processing tasks. From the performance on downstream tasks, we can know that the text representation constructed by these models contains linguistic information, but its encoding mode and extent are unclear. In this work, a structural probe is proposed to detect whether the vector representation produced by a deep neural network is embedded with a syntax tree. The probe is trained with the deep metric learning method, so that the distance between word vectors in the metric space it defines encodes the distance of words on the syntax tree, and the norm of word vectors encodes the depth of words on the syntax tree. The experiment results on ELMo and BERT show that the syntax tree is encoded in their parameters and the word representations they produce.

Keywords: deep metric learning, syntax tree probing, natural language processing, word representations

Procedia PDF Downloads 40

Authors: Theophilus Adedokun, Sylvia Zulu, Felix Awung, Sam Usadolo

Abstract:

The proliferation of technology into teaching process has brought about transformation into the field of education. Language teaching is not left behind from this tremendous transformation which has drastically altered the teaching of language. It is, however, appalling that some language instructors seem to possess negative attitudes toward the use of technology in language teaching, which in this study is referred to as Computer Assisted Language Teaching (CALT). The purpose of this study, therefore, is to explore sustainable lesson that can be learnt from the attitudes of language instructors towards language teaching in some public universities. The knowledge gained from this study could inform and advance the use of Computer Assisted Language Teaching. This study considers the historical progression of CALT and recommends that a fundamental approach is required for institutions to develop and advance the use of CALT for teaching. A review of sustainable lessons learnt from the attitudes of language instructors towards CALT are provided, and the CALT experience of 3 institutions are described. Drawing from this succinct description, this study makes recommendations on how operative CALT could be executed on a personal and institutional basis.

Keywords: attitudes, language instructors, sustainable lessons, computer assisted language teaching

Procedia PDF Downloads 59

Authors: Tahseen Bibi

Abstract:

Language is a source of communicating the ideas, emotions and feelings to others. Languages are different from one another on the basis of symbols and articulation. Regional languages play a role of unification in any country. National language of any country gives strength to its masses as it evaporates the mutual indifferences. There are various regional languages in Pakistan like Sindhi, Pushto, Hindko and Balochi. Hindko language dates back to the ancient times and the Hindko speakers can also easily understand and speak Urdu language. Urdu language is an amalgam of various languages. These languages are interconnected. Thus we can draw an analogy between the two languages under discussion on the basis of the pronunciation. The research will show that there are so many words in both the languages which have the similar pronunciation. It will further tell that the roots of Urdu language lie in Hindko. The reason behind this resemblance is that Urdu has got extracted from Hindko and other languages. Hindko language has played a prominent role in the development of Urdu language. Thus the role of Hindko language in the emergence and development of Urdu cannot be denied. This article will use the qualitative and comparative study as methodology. The research will highlight that there is close resemblance in both the languages on the basis of pronunciation, signifying that Urdu language has been extracted from Hindkon language.

Keywords: Hindko, Urdu, regional languages, vocabulary

Procedia PDF Downloads 379

Authors: Stephen L. Green, Alexander N. Gorban, Ivan Y. Tyukin

Abstract:

Within the past decade, using Convolutional Neural Networks (CNN)’s to create Deep Learning systems capable of translating Sign Language into text has been a breakthrough in breaking the communication barrier for deaf-mute people. Conventional research on this subject has been concerned with training the network to recognize the fingerspelling gestures of a given language and produce their corresponding alphanumerics. One of the problems with the current developing technology is that images are scarce, with little variations in the gestures being presented to the recognition program, often skewed towards single skin tones and hand sizes that makes a percentage of the population’s fingerspelling harder to detect. Along with this, current gesture detection programs are only trained on one finger spelling language despite there being one hundred and forty-two known variants so far. All of this presents a limitation for traditional exploitation for the state of current technologies such as CNN’s, due to their large number of required parameters. This work aims to present a technology that aims to resolve this issue by combining a pretrained legacy AI system for a generic object recognition task with a corrector method to uptrain the legacy network. This is a computationally efficient procedure that does not require large volumes of data even when covering a broad range of sign languages such as American Sign Language, British Sign Language and Chinese Sign Language (Pinyin). Implementing recent results on method concentration, namely the stochastic separation theorem, an AI system is supposed as an operate mapping an input present in the set of images u ∈ U to an output that exists in a set of predicted class labels q ∈ Q of the alphanumeric that q represents and the language it comes from. These inputs and outputs, along with the interval variables z ∈ Z represent the system’s current state which implies a mapping that assigns an element x ∈ ℝⁿ to the triple (u, z, q). As all xi are i.i.d vectors drawn from a product mean distribution, over a period of time the AI generates a large set of measurements xi called S that are grouped into two categories: the correct predictions M and the incorrect predictions Y. Once the network has made its predictions, a corrector can then be applied through centering S and Y by subtracting their means. The data is then regularized by applying the Kaiser rule to the resulting eigenmatrix and then whitened before being split into pairwise, positively correlated clusters. Each of these clusters produces a unique hyperplane and if any element x falls outside the region bounded by these lines then it is reported as an error. As a result of this methodology, a self-correcting recognition process is created that can identify fingerspelling from a variety of sign language and successfully identify the corresponding alphanumeric and what language the gesture originates from which no other neural network has been able to replicate.

Keywords: convolutional neural networks, deep learning, shallow correctors, sign language

Procedia PDF Downloads 81

Authors: Siti Sara binti Hj Ahmad, Adil Elshiekh Abdalla

Abstract:

The Arabic and Malay languages belong to different language’s families; while the Arabic language descends from the Semitic language, Malay belongs to the Austronesian (Malayo-Polynesian) family. Hence, the grammatical systems of the two languages differ from each other. Arabic, being a language found in the heart of the dessert, and Malay is the language found in the heart of thick equatorial forests, is another source of vital cultural differences. Consequently, it is expected that this situation will create differences in the ways of how speakers of the two languages perceive the world around them, convey and understand their messages. On the other hand, as the majority of the speakers of Malay language are Muslims, Arabic language found its way in this region; currently, Arabic is widely taught in school, some terms of it found their way in the Malay language. Accordingly, the Arabic language and culture have widely penetrated into the Malay language. This study is proposed with the aim to find out the differences and similarities between the two languages, in the term of the nonverbal communication. The result of this study will be of high significance, as it will help in enhancing the mutual understanding between the speakers of these languages. The comparative analysis approach will be utilized in this study.

Keywords: gesture, Arabic language, Malay language, comparative analysis

Procedia PDF Downloads 537