Search results

The purpose of this study is to investigate and demonstrate the advancements achieved in the field of chest X-ray image captioning through the utilization of dynamic convolutional encoder–decoder networks (DyCNN). Typical convolutional neural networks (CNNs) are unable to capture both local and global contextual information effectively and apply a uniform operation to all pixels in an image. To address this, we propose an innovative approach that integrates a dynamic convolution operation at the encoder stage, improving image encoding quality and disease detection. In addition, a decoder based on the gated recurrent unit (GRU) is used for language modeling, and an attention network is incorporated to enhance consistency. This novel combination allows for improved feature extraction, mimicking the expertise of radiologists by selectively focusing on important areas and producing coherent captions with valuable clinical information.

In this study, we have presented a new report generation approach that utilizes dynamic convolution applied Resnet-101 (DyCNN) as an encoder (Verelst and Tuytelaars, 2019) and GRU as a decoder (Dey and Salemt, 2017; Pan et al., 2020), along with an attention network (see Figure 1). This integration innovatively extends the capabilities of image encoding and sequential caption generation, representing a shift from conventional CNN architectures. With its ability to dynamically adapt receptive fields, the DyCNN excels at capturing features of varying scales within the CXR images. This dynamic adaptability significantly enhances the granularity of feature extraction, enabling precise representation of localized abnormalities and structural intricacies. By incorporating this flexibility into the encoding process, our model can distil meaningful and contextually rich features from the radiographic data. While the attention mechanism enables the model to selectively focus on different regions of the image during caption generation. The attention mechanism enhances the report generation process by allowing the model to assign different importance weights to different regions of the image, mimicking human perception. In parallel, the GRU-based decoder adds a critical dimension to the process by ensuring a smooth, sequential generation of captions.

The findings of this study highlight the significant advancements achieved in chest X-ray image captioning through the utilization of dynamic convolutional encoder–decoder networks (DyCNN). Experiments conducted using the IU-Chest X-ray datasets showed that the proposed model outperformed other state-of-the-art approaches. The model achieved notable scores, including a BLEU_1 score of 0.591, a BLEU_2 score of 0.347, a BLEU_3 score of 0.277 and a BLEU_4 score of 0.155. These results highlight the efficiency and efficacy of the model in producing precise radiology reports, enhancing image interpretation and clinical decision-making.

This work is the first of its kind, which employs DyCNN as an encoder to extract features from CXR images. In addition, GRU as the decoder for language modeling was utilized and the attention mechanisms into the model architecture were incorporated.

Intent detection (ID) and slot filling (SF) are two important tasks in natural language understanding. ID is to identify the main intent of a paragraph of text. The goal of SF is to extract the information that is important to the intent from the input sentence. However, most of the existing methods use sentence-level intention recognition, which has the risk of error propagation, and the relationship between intention recognition and SF is not explicitly modeled. Aiming at this problem, this paper proposes a collaborative model of ID and SF for intelligent spoken language understanding called ID-SF-Fusion.

ID-SF-Fusion uses Bidirectional Encoder Representation from Transformers (BERT) and Bidirectional Long Short-Term Memory (BiLSTM) to extract effective word embedding and context vectors containing the whole sentence information respectively. Fusion layer is used to provide intent–slot fusion information for SF task. In this way, the relationship between ID and SF task is fully explicitly modeled. This layer takes the result of ID and slot context vectors as input to obtain the fusion information which contains both ID result and slot information. Meanwhile, to further reduce error propagation, we use word-level ID for the ID-SF-Fusion model. Finally, two tasks of ID and SF are realized by joint optimization training.

We conducted experiments on two public datasets, Airline Travel Information Systems (ATIS) and Snips. The results show that the Intent ACC score and Slot F1 score of ID-SF-Fusion on ATIS and Snips are 98.0 per cent and 95.8 per cent, respectively, and the two indicators on Snips dataset are 98.6 per cent and 96.7 per cent, respectively. These models are superior to slot-gated, SF-ID NetWork, stack-Prop and other models. In addition, ablation experiments were performed to further analyze and discuss the proposed model.

This paper uses word-level intent recognition and introduces intent information into the SF process, which is a significant improvement on both data sets.

The paper aims to enhance Arabic machine translation (MT) by proposing novel approaches: (1) a dimensionality reduction technique for word embeddings tailored for Arabic text, optimizing efficiency while retaining semantic information; (2) a comprehensive comparison of meta-embedding techniques to improve translation quality; and (3) a method leveraging self-attention and Gated CNNs to capture token dependencies, including temporal and hierarchical features within sentences, and interactions between different embedding types. These approaches collectively aim to enhance translation quality by combining different embedding schemes and leveraging advanced modeling techniques.

Recent works on MT in general and Arabic MT in particular often pick one type of word embedding model. In this paper, we present a novel approach to enhance Arabic MT by addressing three key aspects. Firstly, we propose a new dimensionality reduction technique for word embeddings, specifically tailored for Arabic text. This technique optimizes the efficiency of embeddings while retaining their semantic information. Secondly, we conduct an extensive comparison of different meta-embedding techniques, exploring the combination of static and contextual embeddings. Through this analysis, we identify the most effective approach to improve translation quality. Lastly, we introduce a novel method that leverages self-attention and Gated convolutional neural networks (CNNs) to capture token dependencies, including temporal and hierarchical features within sentences, as well as interactions between different types of embeddings. Our experimental results demonstrate the effectiveness of our proposed approach in significantly enhancing Arabic MT performance. It outperforms baseline models with a BLEU score increase of 2 points and achieves superior results compared to state-of-the-art approaches, with an average improvement of 4.6 points across all evaluation metrics.

The proposed approaches significantly enhance Arabic MT performance. The dimensionality reduction technique improves the efficiency of word embeddings while preserving semantic information. Comprehensive comparison identifies effective meta-embedding techniques, with the contextualized dynamic meta-embeddings (CDME) model showcasing competitive results. Integration of Gated CNNs with the transformer model surpasses baseline performance, leveraging both architectures' strengths. Overall, these findings demonstrate substantial improvements in translation quality, with a BLEU score increase of 2 points and an average improvement of 4.6 points across all evaluation metrics, outperforming state-of-the-art approaches.

The paper’s originality lies in its departure from simply fine-tuning the transformer model for a specific task. Instead, it introduces modifications to the internal architecture of the transformer, integrating Gated CNNs to enhance translation performance. This departure from traditional fine-tuning approaches demonstrates a novel perspective on model enhancement, offering unique insights into improving translation quality without solely relying on pre-existing architectures. The originality in dimensionality reduction lies in the tailored approach for Arabic text. While dimensionality reduction techniques are not new, the paper introduces a specific method optimized for Arabic word embeddings. By employing independent component analysis (ICA) and a post-processing method, the paper effectively reduces the dimensionality of word embeddings while preserving semantic information which has not been investigated before especially for MT task.

In the era of big data, various industries are generating large amounts of text data every day. Simplifying and summarizing these data can effectively serve users and improve efficiency. Recently, zero-shot prompting in large language models (LLMs) has demonstrated remarkable performance on various language tasks. However, generating a very “concise” multi-document summary is a difficult task for it. When conciseness is specified in the zero-shot prompting, the generated multi-document summary still contains some unimportant information, even with the few-shot prompting. This paper aims to propose a LLMs prompting for multi-document summarization task.

To overcome this challenge, the authors propose chain-of-event (CoE) prompting for multi-document summarization (MDS) task. In this prompting, the authors take events as the center and propose a four-step summary reasoning process: specific event extraction; event abstraction and generalization; common event statistics; and summary generation. To further improve the performance of LLMs, the authors extend CoE prompting with the example of summary reasoning.

Summaries generated by CoE prompting are more abstractive, concise and accurate. The authors evaluate the authors’ proposed prompting on two data sets. The experimental results over ChatGLM2-6b show that the authors’ proposed CoE prompting consistently outperforms other typical promptings across all data sets.

This paper proposes CoE prompting to solve MDS tasks by the LLMs. CoE prompting can not only identify the key events but also ensure the conciseness of the summary. By this method, users can access the most relevant and important information quickly, improving their decision-making processes.

Assessing whether a book is relevant or suitable for use in teaching materials is not an easy and haphazard matter, various methods and theories have been offered by researchers in studying this matter. Taking a study of the context of textbooks, researchers found the urgency that textbooks are a foundation for education, socialization and transmission of knowledge and its construction. Researchers offer another approach, namely by using praxeology as a study tool so that the goals of the textbooks previously intended are fulfilled.

The researcher uses a qualitative approach through grounded theory. Grounded theory procedures are designed to develop a well-integrated set of concepts that provide a thorough theoretical explanation of the social phenomena under study. A grounded theory must explain as well as describe. It may also implicitly provide some degree of predictability, but only with respect to certain conditions (Corbin and Strauss, 1990). Document analysis in conducting this research study. Document analysis itself examines systematic procedures for reviewing or evaluating documents, both printed and electronic materials.

Two issues regarding gender acquisition have been investigated in L2 Arabic acquisition studies; the order in which L2 Arabic learners acquire certain grammatical features of the gender system and the effect of L1 on the acquisition of some grammatical features from L2 grammatical gender. Arabic has a two-gender system that classifies all nouns, animate and inanimate, as masculine or feminine. Verbs, nouns, adjectives, personal, demonstrative and relative pronouns related to nouns in the syntactic structure of sentences show gender agreement.

In practice, as a book intended for non-speakers, the book is presented using a general view of linguistic theory. In relation to the gender agreement, the presentation of the book begins and is inserted with the concepts of nouns and verbs. Returning to the praxeology context, First, The Know How (Praxis) explains practice (i.e. the tasks performed and the techniques used). Second, To Know Why or Knowledge (logos) which explains and justifies practice from a technological and theoretical point of view. Answering the first concept, the exercise presented in the book is a concept with three clusters explained at the beginning of the discussion. And the second concept, explained with a task design approach which includes word categorization by separating masculine and feminine word forms.

Practically, this research obtains perspectives studied from a textbook, namely the Arabic gender agreement is presented with various examples of noun contexts; textbook authors present book concepts in a particular way with regard to curriculum features and this task design affects student performance, and which approach is more effective for developing student understanding. Empirically, the material is in line with the formulation of competency standards for non-Arabic speakers in Indonesia.

With this computational search, the researcher found a novelty that was considered accurate by taking the praxeology context as a review in the analysis of non-speaking Arabic textbooks, especially in the year 2022 (last data collection in September) there has been no study on this context. So then, the researcher finds other interests in that praxeology can examine more broadly parts of the task of the contents of the book with the approach of relevant linguistic theories.

Aspect-based sentiment analysis (ASA) is a task of sentiment analysis that requires predicting aspect sentiment polarity for a given sentence. Many traditional techniques use graph-based mechanisms, which reduce prediction accuracy and introduce large amounts of noise. The other problem with graph-based mechanisms is that for some context words, the feelings change depending on the aspect, and therefore it is impossible to draw conclusions on their own. ASA is challenging because a given sentence can reveal complicated feelings about multiple aspects.

This research proposed an optimized attention-based DL model known as optimized aspect and self-attention aware long short-term memory for target-based semantic analysis (OAS-LSTM-TSA). The proposed model goes through three phases: preprocessing, aspect extraction and classification. Aspect extraction is done using a double-layered convolutional neural network (DL-CNN). The optimized aspect and self-attention embedded LSTM (OAS-LSTM) is used to classify aspect sentiment into three classes: positive, neutral and negative.

To detect and classify sentiment polarity of the aspect using the optimized aspect and self-attention embedded LSTM (OAS-LSTM) model. The results of the proposed method revealed that it achieves a high accuracy of 95.3 per cent for the restaurant dataset and 96.7 per cent for the laptop dataset.

The novelty of the research work is the addition of two effective attention layers in the network model, loss function reduction and accuracy enhancement, using a recent efficient optimization algorithm. The loss function in OAS-LSTM is minimized using the adaptive pelican optimization algorithm, thus increasing the accuracy rate. The performance of the proposed method is validated on four real-time datasets, Rest14, Lap14, Rest15 and Rest16, for various performance metrics.

In the COVID-19 era, sign language (SL) translation has gained attention in online learning, which evaluates the physical gestures of each student and bridges the communication gap between dysphonia and hearing people. The purpose of this paper is to devote the alignment between SL sequence and nature language sequence with high translation performance.

SL can be characterized as joint/bone location information in two-dimensional space over time, forming skeleton sequences. To encode joint, bone and their motion information, we propose a multistream hierarchy network (MHN) along with a vocab prediction network (VPN) and a joint network (JN) with the recurrent neural network transducer. The JN is used to concatenate the sequences encoded by the MHN and VPN and learn their sequence alignments.

We verify the effectiveness of the proposed approach and provide experimental results on three large-scale datasets, which show that translation accuracy is 94.96, 54.52, and 92.88 per cent, and the inference time is 18 and 1.7 times faster than listen-attend-spell network (LAS) and visual hierarchy to lexical sequence network (H2SNet) , respectively.

In this paper, we propose a novel framework that can fuse multimodal input (i.e. joint, bone and their motion stream) and align input streams with nature language. Moreover, the provided framework is improved by the different properties of MHN, VPN and JN. Experimental results on the three datasets demonstrate that our approaches outperform the state-of-the-art methods in terms of translation accuracy and speed.

Aboriginal and Torres Strait Islander (Aboriginal) people are overrepresented in Australian prisons, where they experience complex health needs. A model of care was designed to respond to the broad needs of the Aboriginal prisoner population within the nine adult prisons across South Australia. The purpose of this paper is to describe the methods and findings of the Model of Care for Aboriginal and Torres Strait Islander Prisoner Health and Wellbeing for South Australia.

The project used a qualitative mixed-method approach, including a rapid review of relevant literature, stakeholder consultations and key stakeholder workshop. The project was overseen by a Stakeholder Reference Group, which met monthly to ensure that the specific needs of project partners, stakeholders and Aboriginal communities were appropriately incorporated into the planning and management of the project and to facilitate access to relevant information and key informants.

The model of care for Aboriginal prisoner health and wellbeing is designed to be holistic, person-centred and underpinned by the provision of culturally appropriate care. It recognises that Aboriginal prisoners are members of communities both inside and outside of prison. It notes the unique needs of remanded and sentenced prisoners and differing needs by gender.

Supporting the health and wellbeing of Indigenous prison populations can improve health outcomes, community health and reduce recidivism.

Only one other model of care for Aboriginal prisoner health exists in Australia, an Aboriginal Community Controlled Health Organisation-initiated in-reach model of care in one prison in one jurisdiction. The South Australian model of care presents principles that are applicable across all jurisdictions and provides a framework that could be adapted to support Indigenous peoples in diverse prison settings.

This article presents two Brazilian Portuguese corpora collected from different media concerning public security issues in a specific location. The primary motivation is supporting analyses, so security authorities can make appropriate decisions about their actions.

The corpora were obtained through web scraping from a newspaper's website and tweets from a Brazilian metropolitan region. Natural language processing was applied considering: text cleaning, lemmatization, summarization, part-of-speech and dependencies parsing, named entities recognition, and topic modeling.

Several results were obtained based on the methodology used, highlighting some: an example of a summarization using an automated process; dependency parsing; the most common topics in each corpus; the forty named entities and the most common slogans were extracted, highlighting those linked to public security.

Some critical tasks were identified for the research perspective, related to the applied methodology: the treatment of noise from obtaining news on their source websites, passing through textual elements quite present in social network posts such as abbreviations, emojis/emoticons, and even writing errors; the treatment of subjectivity, to eliminate noise from irony and sarcasm; the search for authentic news of issues within the target domain. All these tasks aim to improve the process to enable interested authorities to perform accurate analyses.

The corpora dedicated to the public security domain enable several analyses, such as mining public opinion on security actions in a given location; understanding criminals' behaviors reported in the news or even on social networks and drawing their attitudes timeline; detecting movements that may cause damage to public property and people welfare through texts from social networks; extracting the history and repercussions of police actions, crossing news with records on social networks; among many other possibilities.

The work on behalf of the corpora reported in this text represents one of the first initiatives to create textual bases in Portuguese, dedicated to Brazil's specific public security domain.

While sustainability experts point to interrelated social, economic and environmental goals, students may think about sustainability primarily as natural resources. To prepare students to tackle global challenges to well-being, this paper aims to show that educators need to assess and address students’ shortcomings in considering socioeconomic dimensions.

This study coded essays on the meaning and components of sustainability written by 93 undergraduate and graduate students in environmental policy, business and engineering courses at US and Austrian universities. Then, the study reviewed a teaching strategy using diverse experts, case studies and assignments. Finally, the analysis evaluated students’ final projects proposing sustainability legislation with social, economic and environmental dimensions.

Students usually connect sustainability with limited natural resources affecting current and future generations, but seldom think that sustainability means acting on prominent socioeconomic challenges like poverty, food insecurity, pandemics and violence. Teaching in diverse courses through multidimensional case studies and legislation broadened and deepened students’ understanding and preparedness to act.

Despite experts’ attention to the interconnected Sustainable Development Goals, educators and policymakers need information on whether students associate sustainability with socioeconomic challenges. Open-response questions can reveal gaps in the respondents’ sustainability beliefs. In a wide range of courses, teaching can use diverse experts and multidimensional case studies and legislative assignments.