The classification of goods involved in international trade in Brazil is based on the Mercosur Common Nomenclature (NCM). The classification of these goods represents a real challenge due to the complexity involved in assigning the correct category codes especially considering the legal and fiscal implications of misclassification. This work focuses on the training of a classifier based on Bidirectional En-coder Representations from Transformers (BERT) for the tax classification of goods with NCM codes. In particular, this article presents results from using a specific Portuguese Language tuned BERT model as well results from using a Multilingual BERT. Experimental results justify the use of these models in the classification process and also that the language specific model has a slightly better performance.

In recent times, BERT based transformer models have become an inseparable part of the 'tech stack' of text processing models. Similar progress is being observed in the speech domain with a multitude of models observing state-of-the-art results by using audio transformer models to encode speech. This begs the question of what are these audio transformer models learning. Moreover, although the standard methodology is to choose the last layer embedding for any downstream task, but is it the optimal choice? We try to answer these questions for the two recent audio transformer models, Mockingjay and wave2vec2.0. We compare them on a comprehensive set of language delivery and structure features including audio, fluency and pronunciation features. Additionally, we probe the audio models' understanding of textual surface, syntax, and semantic features and compare them to BERT. We do this over exhaustive settings for native, non-native, synthetic, read and spontaneous speech datasets

Recently, the focus of entity linking research has centered on the zero-shot scenario, where the entity mention to be labeled at the time of testing was never observed during the training phase or may belong to a different domain than the source domain. Current studies have used BERT as the base encoder as it effectively establishes distributional links between source and target domains. The currently available negative sampling methods all use an extractive approach, which makes it difficult for the models to learn diverse and more challenging negative samples. To address this problem, we propose a generative negative sampling method, Adaptive_mixup_hard, which generates more difficult negative entities by fusing the features of both positive and negative samples on top of hard negative sampling and introduces a transformable adaptive parameter W to increase the diversity of negative samples. Next, we fuse our method with the Biencoder architecture and evaluate its performance under three different score functions. Ultimately, experimental results on the standard benchmark dataset Zeshel demonstrate the effectiveness of our method.

Document classification is one of the most critical steps in the document analysis pipeline. There are two types of approaches for document classification, known as image-based and multimodal approaches. The image-based document classification approaches are solely based on the inherent visual cues of the document images. In contrast, the multimodal approach co-learns the visual and textual features, and it has proved to be more effective. Nonetheless, these approaches require a huge amount of data. This paper presents a novel approach for document classification that works with a small amount of data and outperforms other approaches. The proposed approach incorporates a hierarchical attention network(HAN) for the textual stream and the EfficientNet-B0 for the image stream. The hierarchical attention network in the textual stream uses the dynamic word embedding through fine-tuned BERT. HAN incorporates both the word level and sentence level features. While the earlier approaches rely on training on a large corpus (RVL-CDIP), we show that our approach works with a small amount of data (Tobacco-3482). To this end, we trained the neural network at Tobacco-3428 from scratch. Thereby, we outperform state-of-the-art by obtaining an accuracy of 90.3%. This results in a relative error reduction rate of 7.9%.

Recent advances in deep learning technology have led to increased interest in text-to-image technology, which enables computers to create images from text by simulating the human process of forming mental images. The GAN-based text-to-image technology involves the extraction of features from input text, which are combined with noise and then used as input to a GAN that generates images that are similar to the original images through competition between the generator and discriminator. Although generating images from English text is a mature area of research, text-to-image technology based on multilingualism, such as Korean, is still in its early stages of development. Webtoon is a digital comic format that allows comics to be viewed online. The creation process for webtoons is divided into story planning, content/sketching, coloring, and background drawing. Since each stage of webtoon production requires human intervention, it is both time-consuming and expensive. As a result, deep learning technologies such as automatic coloring and automatic line drawing are being used to reduce human involvement. However, there is a shortage of technology that can assist authors with story creation in webtoon production. Therefore, this study proposes a multilingual text-to-image model capable of generating webtoon images when presented with multilingual input text. The proposed model employs Multilingual BERT to extract feature vectors for multiple languages, and trains a DCGAN in conjunction with the images. The experimental results demonstrate that the model can generate images that are similar to the original images when presented with multilingual input text after training.

In this paper we propose a novel self-supervised approach of keywords and keyphrases retrieval and extraction by an end-to-end deep learning approach, which is trained by contextually self-labelled corpus. Our proposed approach is novel to use contextual and semantic features to extract the keywords and has outperformed the state of the art. Through the experiment the proposed approach has been proved to be better in both semantic meaning and quality than the existing popular algorithms of keyword extraction. In addition, we propose to use contextual features from bidirectional transformers to automatically label short-sentence corpus with keywords and keyphrases to build the ground truth. This process avoids the human time to label the keywords and do not need any prior knowledge. To the best of our knowledge, our published dataset in this paper is a fine domain-independent corpus of short sentences with labelled keywords and keyphrases in the NLP community.

Requirements are predominantly written in Natural Language (NL), which makes them accessible to stakeholders with varying degrees of experience, as compared to a model-based language which requires special training. However, despite its benefits, NL can introduce ambiguities and inconsistencies in requirements, which may eventually result in system quality degradation and system failure altogether. The system complexity that characterizes current systems warrants an integrated and comprehensive approach to system design and development. This need has brought about a paradigm shift towards Model-Based Systems Engineering (MBSE) approaches to system design and a departure from traditional document-centric methods. While, MBSE shows great promise, the ambiguities and inconsistencies present in NL requirements hinder their conversion to models directly. The field of Natural Language Processing (NLP) has demonstrated great potential in facilitating the conversion of NL requirements into a semi-machine-readable format that enables their standardization and use in a model-based environment. A first step towards standardizing requirements consists of classifying them according to the ``type'' (design, functional, performance, etc.) they represent. To that end, a language model capable of classifying requirements needs to be fine-tuned on labeled aerospace requirements. This paper presents the development of an annotated aerospace requirements corpus and its use to fine-tune BERT$_\text{BASE-UNCASED}$ to obtain aeroBERT-Classifier: a new language model for classifying aerospace requirements into design, functional, or performance requirements. A comparison between aeroBERT-Classifier and bart-large-mnli (zero-shot text classification) showcased the superior performance of aeroBERT-Classifier on classifying aerospace requirements despite being fine-tuned on a small labeled dataset.

The scope of a lucrative career promoted by Google through its video distribution platform YouTube 1 has attracted a large number of users to become content creators. An important aspect of this line of work is the feedback received in the form of comments which show how well the content is being received by the audience. However, volume of comments coupled with spam and limited tools for comment classification makes it virtually impossible for a creator to go through each and every comment and gather constructive feedback. Automatic classification of comments is a challenge even for established classification models, since comments are often of variable lengths riddled with slang, symbols and abbreviations. This is a greater challenge where comments are multilingual as the messages are often rife with the respective vernacular. In this work, we have evaluated top-performing classification models and four different vectorizers, for classifying comments which are a mix of different combinations of English and Malayalam (only English, only Malayalam and Mix of English and Malayalam). The statistical analysis of results indicates that Multinomial Naïve Bayes, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Random Forest and Decision Trees offer similar level of accuracy in comment classification. Further, we have also evaluated 3 multilingual sub-types of the novel NLP language model, BERT and compared its’ performance to the conventional machine learning classification techniques. XLM was the top-performing BERT model with an accuracy of 67.31%. Random Forest with Term Frequency Vectorizer was the best the top-performing model out of all the traditional classification models with an accuracy of 63.59%.

The dominance of social media has added to the channels of bullying to perpetrators. Unfortunately, cyberbullying (CB) is the most prevalent phenomenon in today’s cyber world and is a severe threat to the mental and physical health of citizens. This opens the need to develop a robust system to prevent bullying content from online forums, blogs, and social media platforms to manage the impact in our society. Several machine learning (ML) algorithms have been proposed for this purpose. However, their performances are not consistent due to high class imbalance issue and generalisation. In recent years, large language models (LLM) like BERT and RoBERTa have achieved state of the art (SOTA) results in several natural language processing (NLP) tasks. Unfortunately, the LLMs have not been applied extensively. In our paper, we explored the use of these models for cyberbullying (CB) detection. We have prepared a new dataset (D2) from existing studies (Formspring and Twitter). Our experimental results for dataset D1 and D2 showed RoBERTa outperformed other models.

Address parsing is a crucial task in natural language processing, particularly for Chinese addresses. The complex structure and semantic features of Chinese addresses present challenges due to their inherent ambiguity. Additionally, different task scenarios require varying levels of granularity in address components, further complicating the parsing process. To address these challenges and adapt to low-resource environments, we propose CapICL, a novel Chinese address parsing model based on the In-Context Learning (ICL) framework. CapICL leverages a sequence generator, regular expression matching, BERT semantic similarity computation, and GPT modeling to enhance parsing accuracy by incorporating contextual information. We construct the sequence generator using a small annotated dataset, capturing distribution patterns and boundary features of address types to model address structure and semantics, mitigating interference from unnecessary variations. We introduce the REB-KNN algorithm, which selects similar samples for ICL-based parsing using regular expression matching and BERT semantic similarity computation. The selected samples, raw text, and explanatory text are combined to form prompts, and inputted into the GPT model for prediction and address parsing. Experimental results demonstrate significant achievements of CapICL in low-resource environments, reducing dependency on annotated data and computational resources. Our model's effectiveness, adaptability, and broad application potential are validated, showcasing its positive impact in natural language processing and geographical information systems.

Election outcomes have been predicted in the past with the help of various state-of-the-art language models. Sentiment analysis helps in establishing the opinions of the public about a particular subject, a popular experiment known as opinion mining. Twitter has grown in popularity and proven to be a key tool in mining people’s sentiments concerning election and other trending subjects of interest. The outcome of the just concluded Presidential election in Nigeria shifts the focus on Lagos State governorship election. In this study, we propose a Bidirectional Encoder Representations from Transformers (BERT) model for the sentiment analysis of governorship election in Lagos State Nigeria using Twitter data. A total of 800,000 personal and public tweets were scraped from twitter concerning the three prominent contesting candidates using carefully selected search queries. The tweets were preprocessed to avoid noise and inconsistencies. The preprocessed tweets were passed into the pretrained and finetuned BERT model. The result was analyzed to establish the sentiments of the public about the candidates. The social networks of the candidates were also analyzed. The parameter-tuning yield different results with different learning rates (LR). Results showed that the learning rate at 1e-7 gave the best performance and that the smaller the learning rate, the higher the accuracy but the larger the epoch size, the higher the accuracy.

Introduction: Social media platforms such as Facebook, LinkedIn, Twitter, among others have been used as a tool for staging protests, opinion polls, campaign strategy, medium of agitation and a place of interest expression especially during elections. Past studies have established people’s opinion elections using social media posts. The advent of state-of-the-art algorithms for unstructured text processing implies tremendous progress in natural language processing and understanding. Aim: In this work, a Natural Language framework is designed to understand Nigeria 2023 presidential election based on public opinion using Twitter dataset. Methods: Raw datasets concerning discourse around Nigeria 2023 elections from Twitter of 2,059,113 18 dimensions were collected. Sentiment analysis was performed on the preprocessed dataset using three different machine learning models namely: Long Short-Term Memory (LSTM) Recurrent Neural Network, Bidirectional Encoder Representations from Transformers (BERT) and Linear Support Vector Classifier (LSVC) models. Personal tweet analysis of the three candidates provided insight on their campaign strategies and personalities while public tweet analysis established the public’s opinion about them. The performance of the models was also compared using accuracy, recall, false positive rate, precision and F-measure. Results: LSTM model gave an accuracy, precision, recall, AUC and f-measure of 88%, 82.7%, 87.2% , 87.6% and 82.9% respectively; the BERT model gave an accuracy, precision, recall, AUC and f-measure of 94%, 88.5%, 92.5%, 94.7% and 91.7% respectively while the LSVC model gave an accuracy, precision, recall, AUC and f-measure of 73%, 81.4%, 76.4%, 81.2% and 79.2% respectively. Conclusion: The experimental results show that sentiment analysis and other Natural Language Processing tasks can aid in the understanding of the social media space. Results also revealed the leverage of each aspirant towards winning the election. We conclude that sentiment analysis can form a general basis for generating insights for election and modeling election outcomes.

We live in a complex world characterised by complex people, complex times, and complex social, technological, and ecological environments. There is clear evidence that governments are failing at most public matters. The recent COVID-19 pandemic is a high example of global governance failure both at preventing such pandemics and managing the COVID-19 pandemic. It is time that all of us take responsibility and look into ways of collaboratively improving the governance of public matters, our matters. While there are many reasons for government failures, we believe the lack of information availability is a fundamental reason that limits the government’s ability to act smartly and allows the lack of transparency to creep into policy and action leading to corruption and failure. To this end, this paper introduces the concept of deep journalism, a data-driven deep learning-based approach for discovering multi-perspective parameters related to a topic of interest. We build three datasets (a newspaper, a technology magazine, and a Web of Science dataset) and discover the academic, industrial, public, governance, and political parameters for the transportation sector as a case study to introduce deep journalism and our tool DeepJournal (Version 1.0) that implements our proposed approach. We elaborate on 89 transportation parameters and hundreds of dimensions reviewing 400 technical, academic, and news articles. The findings related to the multi-perspective view of transportation reported in this paper show that there are many important problems seen by the public that industry and academia seem to not place their focus on. On the other hand, academia produces much broader and deeper knowledge on the subject such as a wide range of pollutions affecting the people and planet do not get to reach the public eye. Our deep journalism approach could find the gaps and highlight them to the public and other stakeholders.

Long unpunctuated texts containing complex linguistic sentences are a stumbling block to processing any low-resource languages. Thus, approaches that attempt to segment lengthy texts with no proper punctuation into simple candidate sentences are a vitally important preprocessing task in many hard-to-solve NLP applications. In this paper, we propose (PDTS) a punctuation detection approach for segmenting Arabic text, built on top of a multilingual BERT-based model and some generic linguistic rules. Furthermore, we showcase how PDTS can be effectively employed as a text tokenizer for unpunctuated documents (i.e., mimicking the transcribed audio-to-text documents). Experimental findings across two evaluation protocols (involving an ablation study and a human-based judgment) demonstrate that PDTS is practically effective in both performance quality and computational cost.

This research investigates the potential application of generative language models, especially ChatGPT, in aviation safety analysis as a means to enhance the efficiency of safety analyses and accelerate the time it takes to process incident reports. In particular, ChatGPT was leveraged to generate incident synopses from narratives, which were subsequently compared with ground truth synopses from the Aviation Safety Reporting System (ASRS) dataset. The comparison was facilitated by using embeddings from Large Language Models (LLMs), with aeroBERT demonstrating the highest similarity due to its aerospace-specific fine-tuning. A positive correlation was observed between synopsis length and their cosine similarity. In a subsequent phase, human factor issues involved in incidents as identified by ChatGPT were compared to human factor issues identified by safety analysts. A concurrence rate of 61% was found, with ChatGPT demonstrating a cautious approach towards attributing human factor issues. Finally, the model was used to attribute incidents to relevant parties. As no dedicated ground truth column existed for this task, a manual evaluation was conducted. ChatGPT attributed the majority of incidents to the Flight Crew, ATC, Ground Personnel, and Maintenance. This study opens new avenues for leveraging AI in aviation safety analysis.

The increased complexity of modern systems is calling for an integrated and comprehensive approach to system design and development and in particular, a shift towards Model-Based Systems Engineering (MBSE) approaches for system design. The requirements that serve as the foundation for these intricate systems are still primarily expressed in Natural Language (NL), which can contain ambiguities and inconsistencies that hinder their direct translation into models. The colossal developments in the field of Natural Language Processing (NLP) in general and Large Language Models (LLMs) in particular can serve as an enabler for the conversion of NL requirements into semi-machine-readable requirements. This is expected to facilitate their standardization and use in a model-based environment. This paper discusses a two-fold strategy for converting NL requirements into semi-machine-readable requirements using language models. The first approach involves creating a requirements table by extracting information from free-form NL requirements. The second approach is an agile methodology that facilitates the identification of boilerplate templates for different types of requirements based on observed linguistic patterns. For this study, three different LLMs were utilized. Two of these models were fine-tuned versions of Bidirectional Encoder Representations from Transformers (BERT), specifically aeroBERT-NER and aeroBERT-Classifier, which were trained on annotated aerospace corpora. Another LLM, called flair/chunk-english, was utilized to identify sentence chunks present in NL requirements. All three language models were utilized together to achieve the standardization of requirements. To demonstrate the effectiveness of the methodologies, requirements from Parts 23 and 25 of Title 14 Code of Federal Regulations (CFRs) were employed, and a total of two, five, and three boilerplate templates were identified for design, functional, and performance requirements, respectively.

The Global natural and manmade events are exposing the fragility of the tourism industry and its impact on the global economy. Prior to the COVID-19 pandemic, tourism contributed 10.3% to the global GDP and employed 333 million people but saw a significant decline due to the pandemic. Sustainable and smart tourism requires collaboration from all stakeholders and a comprehensive understanding of global and local issues to drive responsible and innovative growth in the sector. This paper presents an approach for leveraging big data and deep learning to dis-cover holistic, multi-perspective (e.g., local, cultural, national, and international) and objective information on a subject. Specifically, we develop a machine learning pipeline to extract parameters from academic literature and public opinions on Twitter, providing a unique and comprehensive view of the industry from both academic and public perspectives. The academic-view dataset was created from the Scopus database and contains 156,759 research articles from 2000 to 2022, which were modelled to identify 33 distinct parameters in 4 categories: Tourism Types, Planning, Challenges, and Media & Technologies. A Twitter dataset of 485,813 tweets was collected over 18 months starting March 2021 to August 2022 to showcase public perception of tourism in Saudi Arabia, which was modelled to reveal 13 parameters categorized into two broader sets: Tourist Attractions and Tourism Services. Discovering system parameters are re-quired to embed autonomous capabilities in systems and for decision-making and problem-solving during system design and operations. The proposed approach improves AI-based information discovery by extending the use of scientific literature, Twitter, and other sources for autonomous, dynamic optimizations of systems, promoting novel research in the tourism sector and contributing to the development of smart and sustainable societies. The paper also presents a comprehensive knowledge structure and literature review of the tourism sector based on over 250 research articles.

Technological advancements and innovations have profoundly changed the lives of people giving rise to smart environments, cities, and societies. As homes are the building block of cities and societies, smart homes are critical to establishing smart living and are expected to play a key role in enabling smart cities and societies. The current academic literature and commercial advancements on smart homes have mainly focused on developing and providing smart functions for homes to provide security management and facilitate the residents in their various activities such as ambiance management. Homes are much more than physical structures, buildings, appliances, operational machines, and systems. Homes are composed of families and are inherently complex phenomena underlined by humans and their relationships with each other, subject to individual, intragroup, intergroup, and intercommunity goals. There is a clear need to understand, define, consolidate existing research, and actualize the overarching roles of smart homes, the roles of smart homes that would serve the needs of future smart cities and societies. This paper introduces our data-driven parameter discovery methodology and uses it to provide, for the first time, an extensive, rather fairly comprehensive, analysis of the families and homes landscape seen through the eyes of academics and the public using over a hundred thousand research papers and nearly a million tweets. We develop a methodology using deep learning, natural language processing (NLP), and big data analytics methods and apply it to automatically discover parameters that capture a comprehensive knowledge and design space of smart families and homes comprising social, political, economic, environmental, and other dimensions. The 66 discovered parameters and the knowledge space comprising 100s of dimensions are explained by reviewing and referencing over 300 articles from the academic literature and tweets. The knowledge and parameters discovered in this paper can be used to develop a holistic understanding of matters related to families and homes facilitating the development of better, community-specific, policies, technologies, solutions, and industries for families and homes, leading to strengthening families and homes, and in turn, empowering sustainable societies across the globe.