Search Results (183)

Automatic detection and recognition of wheel hub text, which can boost the efficiency and accuracy of product information recording, are undermined by the obscurity and orientation variability of text on wheel hubs. To address these issues, this paper constructs a wheel hub text dataset and proposes a wheel hub text detection and recognition model called HubNet. The dataset captured images on real industrial production line scenes, including 446 images, 934 word instances, and 2947 character instances. HubNet is an end-to-end text detection and recognition model, not only comprising conventional detection and recognition heads but also incorporating a feature cross-fusion module, which improves the accuracy of recognizing wheel hub texts by utilizing both global and local features. Experimental results show that on the wheel hub text dataset, the HubNet achieves an accuracy of 86.5%, a recall of 79.4%, and an F1-score of 0.828, and the feature cross-fusion module increases the accuracy by 2% to 4%. The wheel hub dataset and the HubNet offer a significant reference for automatic detection and recognition of wheel hub text. Full article

Energy trade balance has been the main factor behind current account imbalances in many developed and developing countries. This study investigates whether or not renewable energy conveys information to the current account deficit of selected OECD countries. Utilizing a dataset spanning from 1990 to 2021, we apply a Panel Autoregressive Distributed Lag (ARDL) estimator to determine the interrelation of current account deficit (CAB) as a percentage of GDP with selected indicators, namely, net energy import in total final energy consumption (NEI), the share of renewable energy in total electricity production (REN_TEO), and fiscal deficit as a percentage of GDP (FAB). The results of long-term estimations reveal that as net energy import increases, the current account deficit deteriorates. On the other hand, in the case that countries utilize more of renewable energy in their total electricity generation, their current account deficits improve. Thus, we conclude that energy policy matters for the current account balances and subsequently for the well-being of OECD economies. Finally, we find strong evidence for the twin deficit hypothesis, as fiscal deficit is negatively interrelated with current account deficit both in the short-run and long run. In other words, an increase in the level of budget deficit is associated with an upsurge in the current account deficit problem. Furthermore, the Dumitrescu-Hurlin causality test reveals that there is bidirectional heterogeneous causality between current account deficit and budget deficit. Additionally, when the countries in the sample are grouped by their per capita GDP levels, estimations reveal that the direction of interaction between CAB and energy-related indicators (NEI and REN_TEO) does not differ between Group 2 (the ones whose per capita incomes are over USD 25,000 but below USD 50,000) and Group 3 (the ones having more than USD 50,000 per capita income) countries. However, the coefficients of energy-related indicators for Group 2 countries are higher than those of Group 3 ones, suggesting that energy policy matters more for Group 2 countries’ current account imbalances in the long-term. Full article

Planar differential systems whose angular velocity is constant are called rigid or uniform differential systems. The first rigid system goes back to the pendulum clock of Christiaan Huygens in 1656; since then, the interest for the rigid systems has been growing. Thus, at this moment, in MathSciNet there are 108 articles with the words rigid systems or uniform systems in their titles. Here, we study the dynamics of the planar rigid polynomial differential systems with homogeneous nonlinearities of arbitrary degree. More precisely, we characterize the existence and non-existence of limit cycles in this class of rigid systems, and we determine the local phase portraits of their finite and infinite equilibrium points in the Poincaré disc. Finally, we classify the global phase portraits in the Poincaré disc of the rigid polynomial differential systems of degree two, and of one class of rigid polynomial differential systems with cubic homogeneous nonlinearities that can exhibit one limit cycle. Full article

The purpose of representation learning is to encode the entities and relations in a knowledge graph as low-dimensional and real-valued vectors through machine learning technology. Traditional representation learning methods like TransE, a method which models relationships by interpreting them as translations operating on the low-dimensional embeddings of a graph’s entities, are effective for learning the embeddings of knowledge bases, but struggle to effectively model complex relations like one-to-many, many-to-one, and many-to-many. To overcome the above issues, we introduce a new method for knowledge representation, reasoning, and completion based on multi-translation principles and TransE (TransE-MTP). By defining multiple translation principles (MTPs) for different relation types, such as one-to-one and complex relations like one-to-many, many-to-one, and many-to-many, and combining MTPs with a typical translating-based model for modeling multi-relational data (TransE), the proposed method, TransE-MTP, ensures that multiple optimization objectives can be targeted and optimized during training on complex relations, thereby providing superior prediction performance. We implement a prototype of TransE-MTP to demonstrate its effectiveness at link prediction and triplet classification on two prominent knowledge graph datasets: Freebase and Wordnet. Our experimental results show that the proposed method enhanced the performance of both TransE and knowledge graph embedding by translating on hyperplanes (TransH), which confirms its effectiveness and competitiveness. Full article

Running machine learning algorithms for image classification locally on small, cheap, and low-power microcontroller units (MCUs) has advantages in terms of bandwidth, inference time, energy, reliability, and privacy for different applications. Therefore, TinyML focuses on deploying neural networks on MCUs with random access memory sizes between 2 KB and 512 KB and read-only memory storage capacities between 32 KB and 2 MB. Models designed for high-end devices are usually ported to MCUs using model scaling factors provided by the model architecture’s designers. However, our analysis shows that this naive approach of substantially scaling down convolutional neural networks (CNNs) for image classification using such default scaling factors results in suboptimal performance. Consequently, in this paper we present a systematic strategy for efficiently scaling down CNN model architectures to run on MCUs. Moreover, we present our CNN Analyzer, a dashboard-based tool for determining optimal CNN model architecture scaling factors for the downscaling strategy by gaining layer-wise insights into the model architecture scaling factors that drive model size, peak memory, and inference time. Using our strategy, we were able to introduce additional new model architecture scaling factors for MobileNet v1, MobileNet v2, MobileNet v3, and ShuffleNet v2 and to optimize these model architectures. Our best model variation outperforms the MobileNet v1 version provided in the MLPerf Tiny Benchmark on the Visual Wake Words image classification task, reducing the model size by 20.5% while increasing the accuracy by 4.0%. Full article

Internet of Things (IoT) applications and resources are highly vulnerable to flood attacks, including Distributed Denial of Service (DDoS) attacks. These attacks overwhelm the targeted device with numerous network packets, making its resources inaccessible to authorized users. Such attacks may comprise attack references, attack types, sub-categories, host information, malicious scripts, etc. These details assist security professionals in identifying weaknesses, tailoring defense measures, and responding rapidly to possible threats, thereby improving the overall security posture of IoT devices. Developing an intelligent Intrusion Detection System (IDS) is highly complex due to its numerous network features. This study presents an improved IDS for IoT security that employs multimodal big data representation and transfer learning. First, the Packet Capture (PCAP) files are crawled to retrieve the necessary attacks and bytes. Second, Spark-based big data optimization algorithms handle huge volumes of data. Second, a transfer learning approach such as word2vec retrieves semantically-based observed features. Third, an algorithm is developed to convert network bytes into images, and texture features are extracted by configuring an attention-based Residual Network (ResNet). Finally, the trained text and texture features are combined and used as multimodal features to classify various attacks. The proposed method is thoroughly evaluated on three widely used IoT-based datasets: CIC-IoT 2022, CIC-IoT 2023, and Edge-IIoT. The proposed method achieves excellent classification performance, with an accuracy of 98.2%. In addition, we present a game theory-based process to validate the proposed approach formally. Full article

Word sense disambiguation (WSD) remains a persistent challenge in the natural language processing (NLP) community. While various NLP packages exist, the Lesk algorithm in the NLTK library demonstrates suboptimal accuracy. In this research article, we propose an innovative methodology and an open-source framework that effectively addresses the challenges of WSD by optimizing memory usage without compromising accuracy. Our system seamlessly integrates WSD into NLP tasks, offering functionality similar to that provided by the NLTK library. However, we go beyond the existing approaches by introducing a novel idea related to WSD. Specifically, we leverage deep neural networks and consider the language patterns learned by these models as the new gold standard. This approach suggests modifying existing semantic dictionaries, such as WordNet, to align with these patterns. Empirical validation through a series of experiments confirmed the effectiveness of our proposed method, achieving state-of-the-art performance across multiple WSD datasets. Notably, our system does not require the installation of additional software beyond the well-known Python libraries. The classification model is saved in a readily usable text format, and the entire framework (model and data) is publicly available on GitHub for the NLP research community. Full article

In the extensive monitoring of maritime traffic, maritime management frequently encounters incomplete automatic identification system (AIS) data. This deficiency poses significant challenges to safety management, requiring effective methods to infer corresponding ship information. We tackle this issue using a classification approach. Due to the absence of a fixed road network at sea unlike on land, raw trajectories are difficult to convert and cannot be directly fed into neural networks. We devised a latitude–longitude gridding encoding strategy capable of transforming continuous latitude–longitude data into discrete grid points. Simultaneously, we employed a compression algorithm to further extract significant grid points, thereby shortening the encoding sequence. Utilizing natural language processing techniques, we integrate the Word2vec word embedding approach with our novel biLSTM self-attention chunk-max pooling net (biSAMNet) model, enhancing the classification of vessel trajectories. This method classifies targets into ship types and ship lengths within static information. Employing the Taiwan Strait as a case study and benchmarking against CNN, RNN, and methods based on the attention mechanism, our findings underscore our model’s superiority. The biSAMNet achieves an impressive trajectory classification F1 score of 0.94 in the ship category dataset using only five-dimensional word embeddings. Additionally, through ablation experiments, the effectiveness of the Word2vec pre-trained embedding layer is highlighted. This study introduces a novel method for handling ship trajectory data, addressing the challenge of obtaining ship static information when AIS data are unreliable. Full article

Sign language is widely used to facilitate the communication process between deaf people and their surrounding environment. Sign language, like most other languages, is considered a complex language which cannot be mastered easily. Thus, technology can be used as an assistive tool to solve the difficulties and challenges that deaf people face during interactions with society. In this study, an automatic bidirectional translation framework for Arabic Sign Language (ArSL) is designed to assist both deaf and ordinary people to communicate and express themselves easily. Two main modules were intended to translate Arabic sign images into text by utilizing different transfer learning models and to translate the input text into Arabic sign images. A prototype was implemented based on the proposed framework by using several pre-trained convolutional neural network (CNN)-based deep learning models, including the DenseNet121, ResNet152, MobileNetV2, Xception, InceptionV3, NASNetLarge, VGG19, and VGG16 models. A fuzzy string matching score method, as a novel concept, was employed to translate the input text from ordinary people into appropriate sign language images. The dataset was constructed with specific criteria to obtain 7030 images for 14 classes captured from both deaf and ordinary people locally. The prototype was developed to conduct the experiments on the collected ArSL dataset using the utilized CNN deep learning models. The experimental results were evaluated using standard measurement metrics such as accuracy, precision, recall, and F1-score. The performance and efficiency of the ArSL prototype were assessed using a test set of an 80:20 splitting procedure, obtaining accuracy results from the highest to the lowest rates with average classification time in seconds for each utilized model, including (VGG16, 98.65%, 72.5), (MobileNetV2, 98.51%, 100.19), (VGG19, 98.22%, 77.16), (DenseNet121, 98.15%, 80.44), (Xception, 96.44%, 72.54), (NASNetLarge, 96.23%, 84.96), (InceptionV3, 94.31%, 76.98), and (ResNet152, 47.23%, 98.51). The fuzzy matching score is mathematically validated by computing the distance between the input and associative dictionary words. The study results showed the prototype’s ability to successfully translate Arabic sign images into Arabic text and vice versa, with the highest accuracy. This study proves the ability to develop a robust and efficient real-time bidirectional ArSL translation system using deep learning models and the fuzzy string matching score method. Full article

The performance of emotive text classification using affective hierarchical schemes (e.g., WordNet-Affect) is often evaluated using the same traditional measures used to evaluate the performance of when a finite set of isolated classes are used. However, applying such measures means the full characteristics and structure of the emotive hierarchical scheme are not considered. Thus, the overall performance of emotive text classification using emotion hierarchical schemes is often inaccurately reported and may lead to ineffective information retrieval and decision making. This paper provides a comparative investigation into how methods used in hierarchical classification problems in other domains, which extend traditional evaluation metrics to consider the characteristics of the hierarchical classification scheme, can be applied and subsequently improve the classification of emotive texts. This study investigates the classification performance of three widely used classifiers, Naive Bayes, J48 Decision Tree, and SVM, following the application of the aforementioned methods. The results demonstrated that all the methods improved the emotion classification. However, the most notable improvement was recorded when a depth-based method was applied to both the testing and validation data, where the precision, recall, and F1-score were significantly improved by around 70 percentage points for each classifier. Full article

Sign language is a complex language that uses hand gestures, body movements, and facial expressions and is majorly used by the deaf community. Sign language recognition (SLR) is a popular research domain as it provides an efficient and reliable solution to bridge the communication gap between people who are hard of hearing and those with good hearing. Recognizing isolated sign language words from video is a challenging research area in computer vision. This paper proposes a hybrid SLR framework that combines a convolutional neural network (CNN) and an attention-based long-short-term memory (LSTM) neural network. We used MobileNetV2 as a backbone model due to its lightweight structure, which reduces the complexity of the model architecture for deriving meaningful features from the video frame sequence. The spatial features are fed to LSTM optimized with an attention mechanism to select the significant gesture cues from the video frames and focus on salient features from the sequential data. The proposed method is evaluated on a benchmark WLASL dataset with 100 classes based on precision, recall, F1-score, and 5-fold cross-validation metrics. Our methodology acquired an average accuracy of 84.65%. The experiment results illustrate that our model performed effectively and computationally efficiently compared to other state-of-the-art methods. Full article

Analysis of brain signals is essential to the study of mental states and various neurological conditions. The two most prevalent noninvasive signals for measuring brain activities are electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS). EEG, characterized by its higher sampling frequency, captures more temporal features, while fNIRS, with a greater number of channels, provides richer spatial information. Although a few previous studies have explored the use of multimodal deep-learning models to analyze brain activity for both EEG and fNIRS, subject-independent training–testing split analysis remains underexplored. The results of the subject-independent setting directly show the model’s ability on unseen subjects, which is crucial for real-world applications. In this paper, we introduce EF-Net, a new CNN-based multimodal deep-learning model. We evaluate EF-Net on an EEG-fNIRS word generation (WG) dataset on the mental state recognition task, primarily focusing on the subject-independent setting. For completeness, we report results in the subject-dependent and subject-semidependent settings as well. We compare our model with five baseline approaches, including three traditional machine learning methods and two deep learning methods. EF-Net demonstrates superior performance in both accuracy and F1 score, surpassing these baselines. Our model achieves F1 scores of 99.36%, 98.31%, and 65.05% in the subject-dependent, subject-semidependent, and subject-independent settings, respectively, surpassing the best baseline F1 scores by 1.83%, 4.34%, and 2.13% These results highlight EF-Net’s capability to effectively learn and interpret mental states and brain activity across different and unseen subjects. Full article

Specialized languages can activate different sets of semantic features when compared to general language or express concepts through different words according to the domain. The specialized lexicon, i.e., lexical units that denote more specific concepts and knowledge emerging from specific domains, however, co-exists with the common lexicon, i.e., the set of lexical units that denote concepts and knowledge shared by the average speakers, regardless of their specific training or expertise. Communication between specialists and non-specialists can show a big gap between language(s), and therefore lexical units, used by the two groups. However, quite often, semantic and conceptual overlapping between specialized and common lexical units occurs and, in many cases, the specialized and common units refer to close concepts or even point to the same reality. Considering the modeling of meaning in functional lexical resources, this paper puts forth a solution that links common and specialized lexica within the WordNet model framework. We propose a new relation expressing semantic proximity between common and specialized units and define the conditions for its establishment. Besides contributing to the observation and understanding of the process of knowledge specialization and its reflex on the lexicon, the proposed relation allows for the integration of specialized and non-specialized lexicons into a single database, contributing directly to improving communication in specialist/non-specialist contexts, such as teaching–learning situations or health professional-patient interactions, among many others, where code-switching is frequent and necessary. Full article

The identification of compound fault components of a planetary gearbox is especially important for keeping the mechanical equipment working safely. However, the recognition performance of existing deep learning-based methods is limited by insufficient compound fault samples and single label classification principles. To solve the issue, a capsule neural network with an improved feature extractor, named LTSS-BoW-CapsNet, is proposed for the intelligent recognition of compound fault components. Firstly, a feature extractor is constructed to extract fault feature vectors from raw signals, which is based on local temporal self-similarity coupled with bag-of-words models (LTSS-BoW). Then, a multi-label classifier based on a capsule network (CapsNet) is designed, in which the dynamic routing algorithm and average threshold are adopted. The effectiveness of the proposed LTSS-BoW-CapsNet method is validated by processing three compound fault diagnosis tasks. The experimental results demonstrate that our method can via decoupling effectively identify the multi-fault components of different compound fault patterns. The testing accuracy is more than 97%, which is better than the other four traditional classification models. Full article

Housing cetaceans in netted sea pens is not new and is common for many accredited managed-care facilities. Hence, the distinction between sanctuary and sea pen is more about the philosophies of those who run these sanctuary facilities, the effects of these philosophies on the animals’ welfare, and how proponents of these sanctuaries fund the care of these animals. Here, I consider what plans exist for cetacean sanctuaries and discuss the caveats and challenges associated with this form of activist-managed captivity. One goal for stakeholders should be to disregard the emotional connotations of the word “sanctuary” and explore these proposals objectively with the best interest of the animals in mind. Another focus should be related to gauging the public’s understanding of proposed welfare benefits to determine if long-term supporters of donation-based sanctuary models will likely see their expectations met as NGOs and their government partners consider moving forward with cetacean sanctuary experiments. Full article