Search

This paper proposes a hybrid deep learning emotion classification system (HDECS), a hybrid multimodal deep learning system designed for emotion classification in a specific national language. Emotion classification is important in diverse fields, including tailored corporate services, AI advancement, and more. Additionally, most sentiment classification techniques in speaking situations are based on a single modality: voice, conversational text, vital signs, etc. However, analyzing these data presents challenges because of the variations in vocal intonation, text structures, and the impact of external stimuli on physiological signals. Korean poses challenges in natural language processing, including subject omission and spacing issues. To overcome these challenges and enhance emotion classification performance, this paper presents a case study using Korean multimodal data. The case study model involves retraining two pretrained models, LSTM and CNN, until their predictions on the entire dataset reach an agreement rate exceeding 0.75. Predictions are used to generate emotional sentences appended to script data, which are further processed using BERT for final emotion prediction. The research result is evaluated by using categorical cross-entropy (CCE) to measure the difference between the model's predictions and actual labels, F1 score, and accuracy. According to the evaluation, the case model outperforms the existing KLUE/roBERTa model with improvements of 0.5 in CCE, 0.09 in accuracy, and 0.11 in F1 score. As a result, the HDECS is expected to perform well not only on Korean multimodal datasets but also on sentiment classification considering the speech characteristics of various languages and regions. [ABSTRACT FROM AUTHOR]

Text summarization is an information compression technology to extract important information from long text, which has become a challenging research direction in the field of natural language processing. At present, the text summary model based on deep learning has shown good results, but how to more effectively model the relationship between words, more accurately extract feature information and eliminate redundant information is still a problem of concern. This paper proposes a graph neural network model GA-GNN based on gated attention, which effectively improves the accuracy and readability of text summarization. First, the words are encoded using a concatenated sentence encoder to generate a deeper vector containing local and global semantic information. Secondly, the ability to extract key information features is improved by using gated attention units to eliminate local irrelevant information. Finally, the loss function is optimized from the three aspects of contrastive learning, confidence calculation of important sentences, and graph feature extraction to improve the robustness of the model. Experimental validation was conducted on a CNN/Daily Mail dataset and MR dataset, and the results showed that the model in this paper outperformed existing methods. [ABSTRACT FROM AUTHOR]

Humor is a special human expression style, an important "lubricant" for daily communication for people; people can convey emotional messages that are not easily expressed through humor. At present, artificial intelligence is one of the popular research domains; "discourse understanding" is also an important research direction, and how to make computers recognize and understand humorous expressions similar to humans has become one of the popular research domains for natural language processing researchers. In this paper, a humor recognition model (MLSN) based on current humor theory and popular deep learning techniques is proposed for the humor recognition task. The model automatically identifies whether a sentence contains humor expression by capturing the inconsistency, phonetic features, and ambiguity of a joke as semantic features. The model was experimented on three publicly available wisecrack datasets and compared with state-of-the-art language models, and the results demonstrate that the proposed model has better humor recognition accuracy and can contribute to the research on discourse understanding. [ABSTRACT FROM AUTHOR]

Short message services (SMS), microblogging tools, instant message apps, and commercial websites produce numerous short text messages every day. These short text messages are usually guaranteed to reach mass audience with low cost. Spammers take advantage of short texts by sending bulk malicious or unwanted messages. Short texts are difficult to classify because of their shortness, sparsity, rapidness, and informal writing. The effectiveness of the hidden Markov model (HMM) for short text classification has been illustrated in our previous study. However, the HMM has limited capability to handle new words, which are mostly generated by informal writing. In this paper, a hybrid model is proposed to address the informal writing issue by weighting new words for fast short text filtering with high accuracy. The hybrid model consists of an artificial neural network (ANN) and an HMM, which are used for new word weighting and spam filtering, respectively. The weight of a new word is calculated based on the weights of its neighbor, along with the spam and ham (i.e., not spam) probabilities of short text message predicted by the ANN. Performance evaluations on benchmark datasets, including the SMS message data maintained by University of California, Irvine; the movie reviews, and the customer reviews are conducted. The hybrid model operates at a significantly higher speed than deep learning models. The experiment results show that the proposed hybrid model outperforms other prominent machine learning algorithms, achieving a good balance between filtering throughput and accuracy. [ABSTRACT FROM AUTHOR]

Deep learning has become a predominant method for solving data analysis problems in virtually all fields of science and engineering. The increasing complexity and the large volume of data collected by diverse sensor systems have spurred the development of deep learning methods and have fundamentally transformed the way the data are acquired, processed, analyzed, and interpreted. With the rapid development of deep learning technology and its ever-increasing range of successful applications across diverse sensor systems, there is an urgent need to provide a comprehensive investigation of deep learning in this domain from a holistic view. This survey paper aims to contribute to this by systematically investigating deep learning models/methods and their applications across diverse sensor systems. It also provides a comprehensive summary of deep learning implementation tips and links to tutorials, open-source codes, and pretrained models, which can serve as an excellent self-contained reference for deep learning practitioners and those seeking to innovate deep learning in this space. In addition, this paper provides insights into research topics in diverse sensor systems where deep learning has not yet been well-developed, and highlights challenges and future opportunities. This survey serves as a catalyst to accelerate the application and transformation of deep learning in diverse sensor systems. [ABSTRACT FROM AUTHOR]

In image captioning models, the main challenge in describing an image is identifying all the objects by precisely considering the relationships between the objects and producing various captions. Over the past few years, many methods have been proposed, from an attribute-to-attribute comparison approach to handling issues related to semantics and their relationships. Despite the improvements, the existing techniques suffer from inadequate positional and geometrical attributes concepts. The reason is that most of the abovementioned approaches depend on Convolutional Neural Networks (CNNs) for object detection. CNN is notorious for failing to detect equivariance and rotational invariance in objects. Moreover, the pooling layers in CNNs cause valuable information to be lost. Inspired by the recent successful approaches, this paper introduces a novel framework for extracting meaningful descriptions based on a parallelized capsule network that describes the content of images through a high level of understanding of the semantic contents of an image. The main contribution of this paper is proposing a new method that not only overrides the limitations of CNNs but also generates descriptions with a wide variety of words by using Wikipedia. In our framework, capsules focus on the generation of meaningful descriptions with more detailed spatial and geometrical attributes for a given set of images by considering the position of the entities as well as their relationships. Qualitative experiments on the benchmark dataset MS-COCO show that our framework outperforms state-of-the-art image captioning models when describing the semantic content of the images. [ABSTRACT FROM AUTHOR]

With many conveniences afforded by advances in smartphone technology, developing advanced data analysis methods for health-related information from smartphone users has become a fast-growing research topic in the healthcare field. Along these lines, this paper addresses smartphone sensor-based characterization of human motions with neural stochastic differential equations (NSDEs) and a Transformer model. NSDEs and modeling via Transformer networks are two of the most prominent deep learning-based modeling approaches, with significant performance yields in many applications. For the problem of modeling dynamical features, stochastic differential equations and deep neural networks are frequently used paradigms in science and engineering, respectively. Combining these two paradigms in one unified framework has drawn significant interest in the deep learning community, and NSDEs are among the leading technologies for combining these efforts. The use of attention has also become a widely adopted strategy in many deep learning applications, and a Transformer is a deep learning model that uses the mechanism of self-attention. This concept of a self-attention based Transformer was originally introduced for tasks of natural language processing (NLP), and due to its excellent performance and versatility, the scope of its applications is rapidly expanding. By utilizing the techniques of neural stochastic differential equations and a Transformer model along with data obtained from smartphone sensors, we present a deep learning method capable of efficiently characterizing human motions. For characterizing human motions, we encode the high-dimensional sequential data from smartphone sensors into latent variables in a low-dimensional latent space. The concept of the latent variable is particularly useful because it can not only carry condensed information concerning motion data, but also learn their low-dimensional representations. More precisely, we use neural stochastic differential equations for modeling transitions of human motion in a latent space, and rely on a Generative Pre-trained Transformer 2 (GPT2)-based Transformer model for approximating the intractable posterior of conditional latent variables. Our experiments show that the proposed method can yield promising results for the problem of characterizing human motion patterns and some related tasks including user identification. [ABSTRACT FROM AUTHOR]

Organizations and individuals worldwide are becoming increasingly vulnerable to cyberattacks as phishing continues to grow and the number of phishing websites grows. As a result, improved cyber defense necessitates more effective phishing detection (PD). In this paper, we introduce a novel method for detecting phishing sites with high accuracy. Our approach utilizes a Convolution Neural Network (CNN)-based model for precise classification that effectively distinguishes legitimate websites from phishing websites. We evaluate the performance of our model on the PhishTank dataset, which is a widely used dataset for detecting phishing websites based solely on Uniform Resource Locators (URL) features. Our approach presents a unique contribution to the field of phishing detection by achieving high accuracy rates and outperforming previous state-of-the-art models. Experiment results revealed that our proposed method performs well in terms of accuracy and its false-positive rate. We created a real data set by crawling 10,000 phishing URLs from PhishTank and 10,000 legitimate websites and then ran experiments using standard evaluation metrics on the data sets. This approach is founded on integrated and deep learning (DL). The CNN-based model can distinguish phishing websites from legitimate websites with a high degree of accuracy. When binary-categorical loss and the Adam optimizer are used, the accuracy of the k-nearest neighbors (KNN), Natural Language Processing (NLP), Recurrent Neural Network (RNN), and Random Forest (RF) models is 87%, 97.98%, 97.4% and 94.26%, respectively, in contrast to previous publications. Our model outperformed previous works due to several factors, including the use of more layers and larger training sizes, and the extraction of additional features from the PhishTank dataset. Specifically, our proposed model comprises seven layers, starting with the input layer and progressing to the seventh, which incorporates a layer with pooling, convolutional, linear 1 and 2, and linear six layers as the output layers. These design choices contribute to the high accuracy of our model, which achieved a 98.77% accuracy rate. [ABSTRACT FROM AUTHOR]

As a research hotspot in the field of natural language processing (NLP), sentiment analysis can be roughly divided into explicit sentiment analysis and implicit sentiment analysis. However, due to the lack of obvious emotion words in the implicit sentiment analysis task and because the sentiment polarity contained in implicit sentiment words is not easily accurately identified by existing text-processing methods, the implicit sentiment analysis task is one of the most difficult tasks in sentiment analysis. This paper proposes a new preprocessing method for implicit sentiment text classification; this method is named Text To Picture (TTP) in this paper. TTP highlights the sentiment differences between different sentiment polarities in Chinese implicit sentiment text with the help of deep learning by converting original text data into word frequency maps. The differences between sentiment polarities are used as sentiment clues to improve the performance of the Chinese implicit sentiment text classification task. It does this by transforming the original text data into a word frequency map in order to highlight the differences between the sentiment polarities expressed in the implicit sentiment text. We conducted experimental tests on two common datasets (SMP2019, EWECT), and the results show that the accuracy of our method is significantly improved compared with that of the competitor's. On the SMP2019 dataset, the accuracy-improvement range was 4.55–7.06%. On the EWECT dataset, the accuracy was improved by 1.81–3.95%. In conclusion, the new preprocessing method for implicit sentiment text classification proposed in this paper can achieve better classification results. [ABSTRACT FROM AUTHOR]

Fashion is one of the many fields of application that image captioning is being used in. For e-commerce websites holding tens of thousands of images of clothing, automated item descriptions are quite desirable. This paper addresses captioning images of clothing in the Arabic language using deep learning. Image captioning systems are based on Computer Vision and Natural Language Processing techniques because visual and textual understanding is needed for these systems. Many approaches have been proposed to build such systems. The most widely used methods are deep learning methods which use the image model to analyze the visual content of the image, and the language model to generate the caption. Generating the caption in the English language using deep learning algorithms received great attention from many researchers in their research, but there is still a gap in generating the caption in the Arabic language because public datasets are often not available in the Arabic language. In this work, we created an Arabic dataset for captioning images of clothing which we named "ArabicFashionData" because this model is the first model for captioning images of clothing in the Arabic language. Moreover, we classified the attributes of the images of clothing and used them as inputs to the decoder of our image captioning model to enhance Arabic caption quality. In addition, we used the attention mechanism. Our approach achieved a BLEU-1 score of 88.52. The experiment findings are encouraging and suggest that, with a bigger dataset, the attributes-based image captioning model can achieve excellent results for Arabic image captioning. [ABSTRACT FROM AUTHOR]

Understanding actions in videos remains a significant challenge in computer vision, which has been the subject of several pieces of research in the last decades. Convolutional neural networks (CNN) are a significant component of this topic and play a crucial role in the renown of Deep Learning. Inspired by the human vision system, CNN has been applied to visual data exploitation and has solved various challenges in various computer vision tasks and video/image analysis, including action recognition (AR). However, not long ago, along with the achievement of the transformer in natural language processing (NLP), it began to set new trends in vision tasks, which has created a discussion around whether the Vision Transformer models (ViT) will replace CNN in action recognition in video clips. This paper conducts this trending topic in detail, the study of CNN and Transformer for Action Recognition separately and a comparative study of the accuracy-complexity trade-off. Finally, based on the performance analysis's outcome, the question of whether CNN or Vision Transformers will win the race will be discussed. [ABSTRACT FROM AUTHOR]

In today's world, mental health diseases have become highly prevalent, and depression is one of the mental health problems that has become widespread. According to WHO reports, depression is the second-leading cause of the global burden of diseases. In the proliferation of such issues, social media has proven to be a great platform for people to express themselves. Thus, a user's social media can speak a great deal about his/her emotional state and mental health. Considering the high pervasiveness of the disease, this paper presents a novel framework for depression detection from textual data, employing Natural Language Processing and deep learning techniques. For this purpose, a dataset consisting of tweets was created, which were then manually annotated by the domain experts to capture the implicit and explicit depression context. Two variations of the dataset were created, on having binary and one ternary labels, respectively. Ultimately, a deep-learning-based hybrid Sequence, Semantic, Context Learning (SSCL) classification framework with a self-attention mechanism is proposed that utilizes GloVe (pre-trained word embeddings) for feature extraction; LSTM and CNN were used to capture the sequence and semantics of tweets; finally, the GRUs and self-attention mechanism were used, which focus on contextual and implicit information in the tweets. The framework outperformed the existing techniques in detecting the explicit and implicit context, with an accuracy of 97.4 for binary labeled data and 82.9 for ternary labeled data. We further tested our proposed SSCL framework on unseen data (random tweets), for which an F1-score of 94.4 was achieved. Furthermore, in order to showcase the strengths of the proposed framework, we validated it on the "News Headline Data set" for sarcasm detection, considering a dataset from a different domain. It also outmatched the performance of existing techniques in cross-domain validation. [ABSTRACT FROM AUTHOR]

The grading of cancer tissues is still one of the main challenges for pathologists. The development of enhanced analysis strategies hence becomes crucial to accurately identify and further deal with each individual case. Raman spectroscopy (RS) is a promising tool for the classification of tumor tissues as it allows us to obtain the biochemical maps of the tissues under analysis and to observe their evolution in terms of biomolecules, proteins, lipid structures, DNA, vitamins, and so on. However, its potential could be further improved by providing a classification system which would be able to recognize the sample tumor category by taking as input the raw Raman spectroscopy signal; this could provide more reliable responses in shorter time scales and could reduce or eliminate false-positive or -negative diagnoses. Deep Learning techniques have become ubiquitous in recent years, with models able to perform classification with high accuracy in most diverse fields of research, e.g., natural language processing, computer vision, medical imaging. However, deep models often rely on huge labeled datasets to produce reasonable accuracy, otherwise occurring in overfitting issues when the training data is insufficient. In this paper, we propose a chondrogenic tumor CLAssification through wavelet transform of RAman spectra (CLARA), which is able to classify with high accuracy Raman spectra obtained from bone tissues. CLARA recognizes and grades the tumors in the evaluated dataset with 97 % accuracy by exploiting a classification pipeline consisting of the division of the original task in two binary classification steps, where the first is performed on the original RS signals while the latter is accomplished through the use of a hybrid temporal-frequency 2D transform. [ABSTRACT FROM AUTHOR]

Businesses need to use sentiment analysis, powered by artificial intelligence and machine learning to forecast accurately whether or not consumers are satisfied with their offerings. This paper uses a deep learning model to analyze thousands of reviews of Amazon Alexa to predict customer sentiment. The proposed model can be directly applied to any company with an online presence to detect customer sentiment from their reviews automatically. This research aims to present a suitable method for analyzing the users' reviews of Amazon Echo and categorizing them into positive or negative thoughts. A dataset containing reviews of 3150 users has been used in this research work. Initially, a word cloud of positive and negative reviews was plotted, which gave a lot of insight from the text data. After that, a deep learning model using a multinomial naive Bayesian classifier was built and trained using 80% of the dataset. Then the remaining 20% of the dataset was used to test the model. The proposed model gives 93% accuracy. The proposed model has also been compared with four models used in the same domain, outperforming three. [ABSTRACT FROM AUTHOR]

This is a review focused on advances and current limitations of computer vision (CV) and how CV can help us obtain to more autonomous actions in surgery. It is a follow-up article to one that we previously published in Sensors entitled, "Artificial Intelligence Surgery: How Do We Get to Autonomous Actions in Surgery?" As opposed to that article that also discussed issues of machine learning, deep learning and natural language processing, this review will delve deeper into the field of CV. Additionally, non-visual forms of data that can aid computerized robots in the performance of more autonomous actions, such as instrument priors and audio haptics, will also be highlighted. Furthermore, the current existential crisis for surgeons, endoscopists and interventional radiologists regarding more autonomy during procedures will be discussed. In summary, this paper will discuss how to harness the power of CV to keep doctors who do interventions in the loop. [ABSTRACT FROM AUTHOR]

The belt conveyor is an essential piece of equipment in coal mining for coal transportation, and its stable operation is key to efficient production. Belt surface of the conveyor is vulnerable to foreign bodies which can be extremely destructive. In the past decades, much research and numerous approaches to inspect belt status have been proposed, and machine learning-based non-destructive testing (NDT) methods are becoming more and more popular. Deep learning (DL), as a branch of machine learning (ML), has been widely applied in data mining, natural language processing, pattern recognition, image processing, etc. Generative adversarial networks (GAN) are one of the deep learning methods based on generative models and have been proved to be of great potential. In this paper, a novel multi-classification conditional CycleGAN (MCC-CycleGAN) method is proposed to generate and discriminate surface images of damages of conveyor belt. A novel architecture of improved CycleGAN is designed to enhance the classification performance using a limited capacity images dataset. Experimental results show that the proposed deep learning network can generate realistic belt surface images with defects and efficiently classify different damaged images of the conveyor belt surface. [ABSTRACT FROM AUTHOR]

Satellite navigation has become ubiquitous to plan and track travelling. Having access to a vehicle's position enables the prediction of its destination. This opens the possibility to various benefits, such as early warnings of potential hazards, route diversions to pass traffic congestion, and optimizing fuel consumption for hybrid vehicles. Thus, reliably predicting destinations can bring benefits to the transportation industry. This paper investigates using deep learning methods for predicting a vehicle's destination based on its journey history. With this aim, Dense Neural Networks (DNNs), Long Short-Term Memory (LSTM) networks, Bidirectional LSTM (BiLSTM), and networks with and without attention mechanisms are tested. Especially, LSTM and BiLSTM models with attention mechanism are commonly used for natural language processing and text-classification-related applications. On the other hand, this paper demonstrates the viability of these techniques in the automotive and associated industrial domain, aimed at generating industrial impact. The results of using satellite navigation data show that the BiLSTM with an attention mechanism exhibits better prediction performance destination, achieving an average accuracy of 96% against the test set (4% higher than the average accuracy of the standard BiLSTM) and consistently outperforming the other models by maintaining robustness and stability during forecasting. [ABSTRACT FROM AUTHOR]

This article explores the possibilities for federated learning with a deep learning method as a basic approach to train detection models for fake news recognition. Federated learning is the key issue in this research because this kind of learning makes machine learning more secure by training models on decentralized data at decentralized places, for example, at different IoT edges. The data are not transformed between decentralized places, which means that personally identifiable data are not shared. This could increase the security of data from sensors in intelligent houses and medical devices or data from various resources in online spaces. Each station edge could train a model separately on data obtained from its sensors and on data extracted from different sources. Consequently, the models trained on local data on local clients are aggregated at the central ending point. We have designed three different architectures for deep learning as a basis for use within federated learning. The detection models were based on embeddings, CNNs (convolutional neural networks), and LSTM (long short-term memory). The best results were achieved using more LSTM layers (F1 = 0.92). On the other hand, all three architectures achieved similar results. We also analyzed results obtained using federated learning and without it. As a result of the analysis, it was found that the use of federated learning, in which data were decomposed and divided into smaller local datasets, does not significantly reduce the accuracy of the models. [ABSTRACT FROM AUTHOR]

Sentiment prediction remains a challenging and unresolved task in various research fields, including psychology, neuroscience, and computer science. This stems from its high degree of subjectivity and limited input sources that can effectively capture the actual sentiment. This can be even more challenging with only text-based input. Meanwhile, the rise of deep learning and an unprecedented large volume of data have paved the way for artificial intelligence to perform impressively accurate predictions or even human-level reasoning. Drawing inspiration from this, we propose a coverage-based sentiment and subsentence extraction system that estimates a span of input text and recursively feeds this information back to the networks. The predicted subsentence consists of auxiliary information expressing a sentiment. This is an important building block for enabling vivid and epic sentiment delivery (within the scope of this paper) and for other natural language processing tasks such as text summarisation and Q&A. Our approach outperforms the state-of-the-art approaches by a large margin in subsentence prediction (i.e., Average Jaccard scores from 0.72 to 0.89). For the evaluation, we designed rigorous experiments consisting of 24 ablation studies. Finally, our learned lessons are returned to the community by sharing software packages and a public dataset that can reproduce the results presented in this paper. [ABSTRACT FROM AUTHOR]

Transfer learning is a pervasive technology in computer vision and natural language processing fields, yielding exponential performance improvements by leveraging prior knowledge gained from data with different distributions. However, while recent works seek to mature machine learning and deep learning techniques in applications related to wireless communications, a field loosely termed radio frequency machine learning, few have demonstrated the use of transfer learning techniques for yielding performance gains, improved generalization, or to address concerns of training data costs. With modifications to existing transfer learning taxonomies constructed to support transfer learning in other modalities, this paper presents a tailored taxonomy for radio frequency applications, yielding a consistent framework that can be used to compare and contrast existing and future works. This work offers such a taxonomy, discusses the small body of existing works in transfer learning for radio frequency machine learning, and outlines directions where future research is needed to mature the field. [ABSTRACT FROM AUTHOR]

Machine learning (ML) and deep learning (DL) have achieved great success in different tasks. These include computer vision, image segmentation, natural language processing, predicting classification, evaluating time series, and predicting values based on a series of variables. As artificial intelligence progresses, new techniques are being applied to areas like optical spectroscopy and its uses in specific fields, such as the agrifood industry. The performance of ML and DL techniques generally improves with the amount of data available. However, it is not always possible to obtain all the necessary data for creating a robust dataset. In the particular case of agrifood applications, dataset collection is generally constrained to specific periods. Weather conditions can also reduce the possibility to cover the entire range of classifications with the consequent generation of imbalanced datasets. To address this issue, data augmentation (DA) techniques are employed to expand the dataset by adding slightly modified copies of existing data. This leads to a dataset that includes values from laboratory tests, as well as a collection of synthetic data based on the real data. This review work will present the application of DA techniques to optical spectroscopy datasets obtained from real agrifood industry applications. The reviewed methods will describe the use of simple DA techniques, such as duplicating samples with slight changes, as well as the utilization of more complex algorithms based on deep learning generative adversarial networks (GANs), and semi-supervised generative adversarial networks (SGANs). [ABSTRACT FROM AUTHOR]

The technical capabilities of modern Industry 4.0 and Industry 5.0 are vast and growing exponentially daily. The present-day Industrial Internet of Things (IIoT) combines manifold underlying technologies that require real-time interconnection and communication among heterogeneous devices. Smart cities are established with sophisticated designs and control of seamless machine-to-machine (M2M) communication, to optimize resources, costs, performance, and energy distributions. All the sensory devices within a building interact to maintain a sustainable climate for residents and intuitively optimize the energy distribution to optimize energy production. However, this encompasses quite a few challenges for devices that lack a compatible and interoperable design. The conventional solutions are restricted to limited domains or rely on engineers designing and deploying translators for each pair of ontologies. This is a costly process in terms of engineering effort and computational resources. An issue persists that a new device with a different ontology must be integrated into an existing IoT network. We propose a self-learning model that can determine the taxonomy of devices given their ontological meta-data and structural information. The model finds matches between two distinct ontologies using a natural language processing (NLP) approach to learn linguistic contexts. Then, by visualizing the ontological network as a knowledge graph, it is possible to learn the structure of the meta-data and understand the device's message formulation. Finally, the model can align entities of ontological graphs that are similar in context and structure.Furthermore, the model performs dynamic M2M translation without requiring extra engineering or hardware resources. [ABSTRACT FROM AUTHOR]

Hazard detection is fundamental for a safe lunar landing. State-of-the-art autonomous lunar hazard detection relies on 2D image-based and 3D Lidar systems. The lunar south pole is challenging for vision-based methods. The low sun inclination and the terrain rich in topographic features create large areas in shadow, hiding the terrain features. The proposed method utilizes a vision transformer (ViT) model, which is a deep learning architecture based on the transformer blocks used in natural language processing, to solve this problem. Our goal is to train the ViT model to extract terrain features information from low-light RGB images. The results show good performances, especially at high altitudes, beating the UNet, one of the most popular convolutional neural networks, in every scenario. [ABSTRACT FROM AUTHOR]

Machine learning (ML) and especially deep learning (DL) with neural networks have demonstrated an amazing success in all sorts of AI problems, from computer vision to game playing, from natural language processing to speech and image recognition. In many ways, the approach of ML toward solving a class of problems is fundamentally different than the one followed in classical engineering, or with ontologies. While the latter rely on detailed domain knowledge and almost exhaustive search by means of static inference rules, ML adopts the view of collecting large datasets and processes this massive information through a generic learning algorithm that builds up tentative solutions. Combining the capabilities of ontology-based recommendation and ML-based techniques in a hybrid system is thus a natural and promising method to enhance semantic knowledge with statistical models. This merge could alleviate the burden of creating large, narrowly focused ontologies for complicated domains, by using probabilistic or generative models to enhance the predictions without attempting to provide a semantic support for them. In this paper, we present a novel hybrid recommendation system that blends a single architecture of classical knowledge-driven recommendations arising from a tailored ontology with recommendations generated by a data-driven approach, specifically with classifiers and a neural collaborative filtering. We show that bringing together these knowledge-driven and data-driven worlds provides some measurable improvement, enabling the transfer of semantic information to ML and, in the opposite direction, statistical knowledge to the ontology. Moreover, the novel proposed system enables the extraction of the reasoning recommendation results after updating the standard ontology with the new products and user behaviors, thus capturing the dynamic behavior of the environment of our interest. [ABSTRACT FROM AUTHOR]

Emotion recognition, among other natural language processing tasks, has greatly benefited from the use of large transformer models. Deploying these models on resource-constrained devices, however, is a major challenge due to their computational cost. In this paper, we show that the combination of large transformers, as high-quality feature extractors, and simple hardware-friendly classifiers based on linear separators can achieve competitive performance while allowing real-time inference and fast training. Various solutions including batch and Online Sequential Learning are analyzed. Additionally, our experiments show that latency and performance can be further improved via dimensionality reduction and pre-training, respectively. The resulting system is implemented on two types of edge device, namely an edge accelerator and two smartphones. [ABSTRACT FROM AUTHOR]

Certain difficulties in path forecasting and filtering problems are based in the initial hypothesis of estimation and filtering techniques. Common hypotheses include that the system can be modeled as linear, Markovian, Gaussian, or all at one time. Although, in many cases, there are strategies to tackle problems with approaches that show very good results, the associated engineering process can become highly complex, requiring a great deal of time or even becoming unapproachable. To have tools to tackle complex problems without starting from a previous hypothesis but to continue to solve classic challenges and sharpen the implementation of estimation and filtering systems is of high scientific interest. This paper addresses the forecast–filter problem from deep learning paradigms with a neural network architecture inspired by natural language processing techniques and data structure. Unlike Kalman, this proposal performs the process of prediction and filtering in the same phase, while Kalman requires two phases. We propose three different study cases of incremental conceptual difficulty. The experimentation is divided into five parts: the standardization effect in raw data, proposal validation, filtering, loss of measurements (forecasting), and, finally, robustness. The results are compared with a Kalman filter, showing that the proposal is comparable in terms of the error within the linear case, with improved performance when facing non-linear systems. [ABSTRACT FROM AUTHOR]

Hyperspectral image (HSI) classification is the subject of intense research in remote sensing. The tremendous success of deep learning in computer vision has recently sparked the interest in applying deep learning in hyperspectral image classification. However, most deep learning methods for hyperspectral image classification are based on convolutional neural networks (CNN). Those methods require heavy GPU memory resources and run time. Recently, another deep learning model, the transformer, has been applied for image recognition, and the study result demonstrates the great potential of the transformer network for computer vision tasks. In this paper, we propose a model for hyperspectral image classification based on the transformer, which is widely used in natural language processing. Besides, we believe we are the first to combine the metric learning and the transformer model in hyperspectral image classification. Moreover, to improve the model classification performance when the available training samples are limited, we use the 1-D convolution and Mish activation function. The experimental results on three widely used hyperspectral image data sets demonstrate the proposed model's advantages in accuracy, GPU memory cost, and running time. [ABSTRACT FROM AUTHOR]

Social media applications, such as Twitter and Facebook, allow users to communicate and share their thoughts, status updates, opinions, photographs, and videos around the globe. Unfortunately, some people utilize these platforms to disseminate hate speech and abusive language. The growth of hate speech may result in hate crimes, cyber violence, and substantial harm to cyberspace, physical security, and social safety. As a result, hate speech detection is a critical issue for both cyberspace and physical society, necessitating the development of a robust application capable of detecting and combating it in real-time. Hate speech detection is a context-dependent problem that requires context-aware mechanisms for resolution. In this study, we employed a transformer-based model for Roman Urdu hate speech classification due to its ability to capture the text context. In addition, we developed the first Roman Urdu pre-trained BERT model, which we named BERT-RU. For this purpose, we exploited the capabilities of BERT by training it from scratch on the largest Roman Urdu dataset consisting of 173,714 text messages. Traditional and deep learning models were used as baseline models, including LSTM, BiLSTM, BiLSTM + Attention Layer, and CNN. We also investigated the concept of transfer learning by using pre-trained BERT embeddings in conjunction with deep learning models. The performance of each model was evaluated in terms of accuracy, precision, recall, and F-measure. The generalization of each model was evaluated on a cross-domain dataset. The experimental results revealed that the transformer-based model, when directly applied to the classification task of the Roman Urdu hate speech, outperformed traditional machine learning, deep learning models, and pre-trained transformer-based models in terms of accuracy, precision, recall, and F-measure, with scores of 96.70%, 97.25%, 96.74%, and 97.89%, respectively. In addition, the transformer-based model exhibited superior generalization on a cross-domain dataset. [ABSTRACT FROM AUTHOR]

Recent advances with large-scale pre-trained language models (e.g., BERT) have brought significant potential to natural language processing. However, the large model size hinders their use in IoT and edge devices. Several studies have utilized task-specific knowledge distillation to compress the pre-trained language models. However, to reduce the number of layers in a large model, a sound strategy for distilling knowledge to a student model with fewer layers than the teacher model is lacking. In this work, we present Layer-wise Adaptive Distillation (LAD), a task-specific distillation framework that can be used to reduce the model size of BERT. We design an iterative aggregation mechanism with multiple gate blocks in LAD to adaptively distill layer-wise internal knowledge from the teacher model to the student model. The proposed method enables an effective knowledge transfer process for a student model, without skipping any teacher layers. The experimental results show that both the six-layer and four-layer LAD student models outperform previous task-specific distillation approaches during GLUE tasks. [ABSTRACT FROM AUTHOR]

At present, the state-of-the-art approaches of Visual Question Answering (VQA) mainly use the co-attention model to relate each visual object with text objects, which can achieve the coarse interactions between multimodalities. However, they ignore the dense self-attention within question modality. In order to solve this problem and improve the accuracy of VQA tasks, in the present paper, an effective Dense Co-Attention Networks (DCAN) is proposed. First, to better capture the relationship between words that are relatively far apart and make the extracted semantics more robust, the Bidirectional Long Short-Term Memory (Bi-LSTM) neural network is introduced to encode questions and answers; second, to realize the fine-grained interactions between the question words and image regions, a dense multimodal co-attention model is proposed. The model's basic components include the self-attention unit and the guided-attention unit, which are cascaded in depth to form a hierarchical structure. The experimental results on the VQA-v2 dataset show that DCAN has obvious performance advantages, which makes VQA applicable to a wider range of AI scenarios. [ABSTRACT FROM AUTHOR]

The concept of the Internet of Medical Things brings a promising option to utilize various electronic health records stored in different medical devices and servers to create practical but secure clinical decision support systems. To achieve such a system, we need to focus on several aspects, most notably the usability aspect of deploying it using low-end devices. This study introduces one such application, namely FedSepsis, for the early detection of sepsis using electronic health records. We incorporate several cutting-edge deep learning techniques for the prediction and natural-language processing tasks. We also explore the multimodality aspect for the better use of electronic health records. A secure distributed machine learning mechanism is essential to building such a practical internet of medical things application. To address this, we analyze two federated learning techniques. Moreover, we use two different kinds of low-computational edge devices, namely Raspberry Pi and Jetson Nano, to address the challenges of using such a system in a practical setting and report the comparisons. We report several critical system-level information about the devices, namely CPU utilization, disk utilization, process CPU threads in use, process memory in use (non-swap), process memory available (non-swap), system memory utilization, temperature, and network traffic. We publish the prediction results with the evaluation metrics area under the receiver operating characteristic curve, the area under the precision–recall curve, and the earliness to predict sepsis in hours. Our results show that the performance is satisfactory, and with a moderate amount of devices, the federated learning setting results are similar to the single server-centric setting. Multimodality provides the best results compared to any single modality in the input features obtained from the electronic health records. Generative adversarial neural networks provide a clear superiority in handling the sparsity of electronic health records. Multimodality with the generative adversarial neural networks provides the best result: the area under the precision–recall curve is 96.55%, the area under the receiver operating characteristic curve is 99.35%, and earliness is 4.56 h. FedSepsis suggests that incorporating such a concept together with low-end computational devices could be beneficial for all the medical sector stakeholders and should be explored further. [ABSTRACT FROM AUTHOR]

The accurate severity classification of a bug report is an important aspect of bug fixing. The bug reports are submitted into the bug tracking system with high speed, and owing to this, bug repository size has been increasing at an enormous rate. This increased bug repository size introduces biases in the bug triage process. Therefore, it is necessary to classify the severity of a bug report to balance the bug triaging process. Previously, many machine learning models were proposed for automation of bug severity classification. The accuracy of these models is not up to the mark because they do not extract the important feature patterns for learning the classifier. This paper proposes a novel deep learning model for multiclass severity classification called Bug Severity classification to address these challenges by using a Convolutional Neural Network and Random forest with Boosting (BCR). This model directly learns the latent and highly representative features. Initially, the natural language techniques preprocess the bug report text, and then n-gram is used to extract the features. Further, the Convolutional Neural Network extracts the important feature patterns of respective severity classes. Lastly, the random forest with boosting classifies the multiple bug severity classes. The average accuracy of the proposed model is 96.34% on multiclass severity of five open source projects. The average F-measures of the proposed BCR and the existing approach were 96.43% and 84.24%, respectively, on binary class severity classification. The results prove that the proposed BCR approach enhances the performance of bug severity classification over the state-of-the-art techniques. [ABSTRACT FROM AUTHOR]

The convolutional neural network (CNN) has become a powerful tool in machine learning (ML) that is used to solve complex problems such as image recognition, natural language processing, and video analysis. Notably, the idea of exploring convolutional neural network architecture has gained substantial attention as well as popularity. This study focuses on the intrinsic various CNN architectures: LeNet, AlexNet, VGG16, ResNet-50, and Inception-V1, which have been scrutinized and compared with each other for the detection of lung cancer using publicly available LUNA16 datasets. Furthermore, multiple performance optimizers: root mean square propagation (RMSProp), adaptive moment estimation (Adam), and stochastic gradient descent (SGD), were applied for this comparative study. The performances of the three CNN architectures were measured for accuracy, specificity, sensitivity, positive predictive value, false omission rate, negative predictive value, and F1 score. The experimental results showed that the CNN AlexNet architecture with the SGD optimizer achieved the highest validation accuracy for CT lung cancer with an accuracy of 97.42%, misclassification rate of 2.58%, 97.58% sensitivity, 97.25% specificity, 97.58% positive predictive value, 97.25% negative predictive value, false omission rate of 2.75%, and F1 score of 97.58%. AlexNet with the SGD optimizer was the best and outperformed compared to the other state-of-the-art CNN architectures. [ABSTRACT FROM AUTHOR]

The growth of the Internet has expanded the amount of data expressed by users across multiple platforms. The availability of these different worldviews and individuals' emotions empowers sentiment analysis. However, sentiment analysis becomes even more challenging due to a scarcity of standardized labeled data in the Bangla NLP domain. The majority of the existing Bangla research has relied on models of deep learning that significantly focus on context-independent word embeddings, such as Word2Vec, GloVe, and fastText, in which each word has a fixed representation irrespective of its context. Meanwhile, context-based pre-trained language models such as BERT have recently revolutionized the state of natural language processing. In this work, we utilized BERT's transfer learning ability to a deep integrated model CNN-BiLSTM for enhanced performance of decision-making in sentiment analysis. In addition, we also introduced the ability of transfer learning to classical machine learning algorithms for the performance comparison of CNN-BiLSTM. Additionally, we explore various word embedding techniques, such as Word2Vec, GloVe, and fastText, and compare their performance to the BERT transfer learning strategy. As a result, we have shown a state-of-the-art binary classification performance for Bangla sentiment analysis that significantly outperforms all embedding and algorithms. [ABSTRACT FROM AUTHOR]

IoT plays an important role in daily life; commands and data transfer rapidly between the servers and objects to provide services. However, cyber threats have become a critical factor, especially for IoT servers. There should be a vigorous way to protect the network infrastructures from various attacks. IDS (Intrusion Detection System) is the invisible guardian for IoT servers. Many machine learning methods have been applied in IDS. However, there is a need to improve the IDS system for both accuracy and performance. Deep learning is a promising technique that has been used in many areas, including pattern recognition, natural language processing, etc. The deep learning reveals more potential than traditional machine learning methods. In this paper, sequential model is the key point, and new methods are proposed by the features of the model. The model can collect features from the network layer via tcpdump packets and application layer via system routines. Text-CNN and GRU methods are chosen because the can treat sequential data as a language model. The advantage compared with the traditional methods is that they can extract more features from the data and the experiments show that the deep learning methods have higher F1-score. We conclude that the sequential model-based intrusion detection system using deep learning method can contribute to the security of the IoT servers. [ABSTRACT FROM AUTHOR]

Computing devices that can recognize various human activities or movements can be used to assist people in healthcare, sports, or human–robot interaction. Readily available data for this purpose can be obtained from the accelerometer and the gyroscope built into everyday smartphones. Effective classification of real-time activity data is, therefore, actively pursued using various machine learning methods. In this study, the transformer model, a deep learning neural network model developed primarily for the natural language processing and vision tasks, was adapted for a time-series analysis of motion signals. The self-attention mechanism inherent in the transformer, which expresses individual dependencies between signal values within a time series, can match the performance of state-of-the-art convolutional neural networks with long short-term memory. The performance of the proposed adapted transformer method was tested on the largest available public dataset of smartphone motion sensor data covering a wide range of activities, and obtained an average identification accuracy of 99.2% as compared with 89.67% achieved on the same data by a conventional machine learning method. The results suggest the expected future relevance of the transformer model for human activity recognition. [ABSTRACT FROM AUTHOR]

Successful applications of deep learning technologies in the natural language processing domain have improved text-based intent classifications. However, in practical spoken dialogue applications, the users' articulation styles and background noises cause automatic speech recognition (ASR) errors, and these may lead language models to misclassify users' intents. To overcome the limited performance of the intent classification task in the spoken dialogue system, we propose a novel approach that jointly uses both recognized text obtained by the ASR model and a given labeled text. In the evaluation phase, only the fine-tuned recognized language model (RLM) is used. The experimental results show that the proposed scheme is effective at classifying intents in the spoken dialogue system containing ASR errors. [ABSTRACT FROM AUTHOR]

The production and consumption of music in the contemporary era results in big data generation and creates new needs for automated and more effective management of these data. Automated music mood detection constitutes an active task in the field of MIR (Music Information Retrieval). The first approach to correlating music and mood was made in 1990 by Gordon Burner who researched the way that musical emotion affects marketing. In 2016, Lidy and Schiner trained a CNN for the task of genre and mood classification based on audio. In 2018, Delbouys et al. developed a multi-modal Deep Learning system combining CNN and LSTM architectures and concluded that multi-modal approaches overcome single channel models. This work will examine and compare single channel and multi-modal approaches for the task of music mood detection applying Deep Learning architectures. Our first approach tries to utilize the audio signal and the lyrics of a musical track separately, while the second approach applies a uniform multi-modal analysis to classify the given data into mood classes. The available data we will use to train and evaluate our models comes from the MoodyLyrics dataset, which includes 2000 song titles with labels from four mood classes, {happy, angry, sad, relaxed}. The result of this work leads to a uniform prediction of the mood that represents a music track and has usage in many applications. [ABSTRACT FROM AUTHOR]

Recommender systems have been applied in a wide range of domains such as e-commerce, media, banking, and utilities. This kind of system provides personalized suggestions based on large amounts of data to increase user satisfaction. These suggestions help client select products, while organizations can increase the consumption of a product. In the case of social data, sentiment analysis can help gain better understanding of a user's attitudes, opinions and emotions, which is beneficial to integrate in recommender systems for achieving higher recommendation reliability. On the one hand, this information can be used to complement explicit ratings given to products by users. On the other hand, sentiment analysis of items that can be derived from online news services, blogs, social media or even from the recommender systems themselves is seen as capable of providing better recommendations to users. In this study, we present and evaluate a recommendation approach that integrates sentiment analysis into collaborative filtering methods. The recommender system proposal is based on an adaptive architecture, which includes improved techniques for feature extraction and deep learning models based on sentiment analysis. The results of the empirical study performed with two popular datasets show that sentiment–based deep learning models and collaborative filtering methods can significantly improve the recommender system's performance. [ABSTRACT FROM AUTHOR]

As COVID-19 solidifies its presence in everyday life, the interest in mental health is growing, resulting in the necessity of sentiment analysis. A smart mirror is suitable for encouraging mental comfort due to its approachability and scalability as an in-home AI device. From the aspect of natural language processing (NLP), sentiment analysis for Korean lacks an emotion dataset regarding everyday conversation. Its significant differences from English in terms of language structure make implementation challenging. The proposed smart mirror LUX provides Korean text sentiment analysis with the deep learning model, which examines GRU, LSTM, CNN, Bi-LSTM, and Bi-GRU networks. There are four emotional labels: anger, sadness, neutral, and happiness. For each emotion, there are three possible interactive responses: reciting wise sayings, playing music, and sympathizing. The implemented smart mirror also includes more-typical functions, such as a wake-up prompt, a weather reporting function, a calendar, a news reporting function, and a clock. [ABSTRACT FROM AUTHOR]

Real-word errors are characterized by being actual terms in the dictionary. By providing context, real-word errors are detected. Traditional methods to detect and correct such errors are mostly based on counting the frequency of short word sequences in a corpus. Then, the probability of a word being a real-word error is computed. On the other hand, state-of-the-art approaches make use of deep learning models to learn context by extracting semantic features from text. In this work, a deep learning model were implemented for correcting real-word errors in clinical text. Specifically, a Seq2seq Neural Machine Translation Model mapped erroneous sentences to correct them. For that, different types of error were generated in correct sentences by using rules. Different Seq2seq models were trained and evaluated on two corpora: the Wikicorpus and a collection of three clinical datasets. The medicine corpus was much smaller than the Wikicorpus due to privacy issues when dealing with patient information. Moreover, GloVe and Word2Vec pretrained word embeddings were used to study their performance. Despite the medicine corpus being much smaller than the Wikicorpus, Seq2seq models trained on the medicine corpus performed better than those models trained on the Wikicorpus. Nevertheless, a larger amount of clinical text is required to improve the results. [ABSTRACT FROM AUTHOR]

Deep Learning (DL) is the state-of-the-art machine learning technology, which shows superior performance in computer vision, bioinformatics, natural language processing, and other areas. Especially as a modern image processing technology, DL has been successfully applied in various tasks, such as object detection, semantic segmentation, and scene analysis. However, with the increase of dense scenes in reality, due to severe occlusions, and small size of objects, the analysis of dense scenes becomes particularly challenging. To overcome these problems, DL recently has been increasingly applied to dense scenes and has begun to be used in dense agricultural scenes. The purpose of this review is to explore the applications of DL for dense scenes analysis in agriculture. In order to better elaborate the topic, we first describe the types of dense scenes in agriculture, as well as the challenges. Next, we introduce various popular deep neural networks used in these dense scenes. Then, the applications of these structures in various agricultural tasks are comprehensively introduced in this review, including recognition and classification, detection, counting and yield estimation. Finally, the surveyed DL applications, limitations and the future work for analysis of dense images in agriculture are summarized. [ABSTRACT FROM AUTHOR]