Search

Air writing is a unique method that allows individuals to freely write or draw in the air without the need for sensor devices. One of the significant applications is aimed at addressing the challenges faced by dyslexic children, who often struggle with reading, writing, and spelling. Currently, there is a lack of tailored applications to guide dyslexic children through their learning process. This paper presents a groundbreaking solution designed to assist dyslexic individuals in mastering the art of writing alphabets, digits, and words. Using various hand gestures, children can effortlessly write any character, digit, or word on a virtual canvas and memorize them easily through the hand movements. This approach not only makes learning enjoyable but also boosts a child's confidence. Additionally, the application allows children to unleash their imagination by creating doodles. While existing air writing methods have limitations, such as reliance on devices and restricting content to digits and characters within predefined boundaries. These limitations have been overcome in our proposed approach. Leveraging afordable technologies, our system caters to the needs of dyslexic learners and serves as an interactive tool to enhance their writing skills. It also provides a means of communication for the deaf and mute community. Utilizing cutting-edge technologies such as machine learning, deep learning, computer vision, and web development, we have developed a practical application that seamlessly tracks finger movements and gestures, capturing trajectories and images for classification. Our model exhibits remarkable performance, achieving a training accuracy rate of 98.57% for digit recognition, 98.80% for character recognition, and 97.09% for doodle recognition. This paper contributes valuable insights into the development of an effective and engaging tool to improve learning skills for students, particularly dyslexic children, while also addressing communication barriers for the deaf and mute community. Our approach opens up new possibilities for inclusive and accessible learning experiences. [ABSTRACT FROM AUTHOR]

Large deep learning models are impressive, but they struggle when real-time data is not available. Few-shot class-incremental learning (FSCIL) poses a significant challenge for deep neural networks to learn new tasks from just a few labeled samples without forgetting the previously learned ones. This setup can easily leads to catastrophic forgetting and overfitting problems, severely affecting model performance. Studying FSCIL helps overcome deep learning model limitations on data volume and acquisition time, while improving practicality and adaptability of machine learning models. This paper provides a comprehensive survey on FSCIL. Unlike previous surveys, we aim to synthesize few-shot learning and incremental learning, focusing on introducing FSCIL from two perspectives, while reviewing over 30 theoretical research studies and more than 20 applied research studies. From the theoretical perspective, we provide a novel categorization approach that divides the field into five subcategories, including traditional machine learning methods, meta learning-based methods, feature and feature space-based methods, replay-based methods, and dynamic network structure-based methods. We also evaluate the performance of recent theoretical research on benchmark datasets of FSCIL. From the application perspective, FSCIL has achieved impressive achievements in various fields of computer vision such as image classification, object detection, and image segmentation, as well as in natural language processing and graph. We summarize the important applications. Finally, we point out potential future research directions, including applications, problem setups, and theory development. Overall, this paper offers a comprehensive analysis of the latest advances in FSCIL from a methodological, performance, and application perspective. • In-depth survey on FSCIL methods, applications, performance. • Categorizing FSCIL into five subcategories for clear analysis. • Evaluating FSCIL research on benchmarks for strengths, weaknesses. • FSCIL applications in computer vision, NLP, and graph. [ABSTRACT FROM AUTHOR]

For real-life image classification problems, publicly available pretrained deep convolutional neural networks (CNN) are often applied. Such families of architectures as ResNet, VGG, or Inception are usually the first choice because they are widely used and can provide satisfying results. Though universal, such approaches can be too sophisticated and computationally expensive when fitting to certain medical image classification tasks, especially those where images are not varied and do not contain features learnt by the models in the pretraining phase. The question of whether some simplified models, dedicated to such tasks, can be designed is of great concern. We performed a case study, using three datasets from the MedMNIST data collection, to test neural network classifiers with simple architectures. This paper presents simple convolutional neural network architectures that perform for the considered data sets similarly to much more sophisticated pre-trained models and automatic ML environments. [ABSTRACT FROM AUTHOR]

The image-based automatic plant disease identification has emerged as a prominent research area, driven by recent advancements in machine learning algorithms. Conventional disease detection methods rely on manual feature extraction from acquired images, involving extensive expertise from experts and pathologists. This labor-intensive practice is often time-consuming and impractical. With the recent advancements in digital technologies and computer vision, Deep neural networks address these challenges by learning the features automatically. This paper focuses on the EfficientNet B0 architecture-based deep neural network model integrated with a convolutional block attention mechanism module for identifying major wheat crop diseases. The attention module in the EfficientNet B0 model improves its performance by emphasizing the disease-infected regions in leaf images. To evaluate the proposed methodology, experiments were conducted on the WheatRust21 image dataset, comprising 6556 images of three major fungal diseases, namely, stripe rust, leaf rust, and stem rust, collected in natural field conditions from 2019 to 2021. Our developed EfficientNet B0-CBAM model performed well with a testing accuracy of 98.70% and an impressive F1 score of 98% on the dataset. The Grad-CAM visualization demonstrates that the attention module focuses more on the infected region and plays a crucial role in wheat disease identification. The results are quantitatively substantial, with the model achieving high accuracy and other evaluation metrics scores with an average of 98.5 %. These findings hold promise for practical applications in agriculture and sustainable crop management. [ABSTRACT FROM AUTHOR]

Multiple Instance Learning (MIL) is a learning paradigm, where training instances are arranged in sets, called bags, and only bag-level labels are available during training. This learning paradigm has been successfully applied in various real-world scenarios, including medical image analysis, object detection, image classification, drug activity prediction, and many others. This survey paper presents a comprehensive analysis of MIL, highlighting its significance, recent advancements, methodologies, applications, and evolving trends across diverse domains. The survey begins by explaining the core principles that form the basis of MIL and how it differs from traditional learning approaches. This sets the foundation for comprehending the distinct challenges and techniques of solving MIL problems. Next, we discuss how supervised learning algorithms are tailored to support MIL and combine this discussion with a review of seminal MIL algorithms as well as the latest innovations that incorporate neural networks, deep learning architectures, and attention techniques. This comprehensive analysis helps to understand the strengths, limitations, and adaptability of these methods across diverse data modalities, complexities, and applications. In summary, this survey paper provides an essential resource for researchers, practitioners, and enthusiasts seeking a comprehensive understanding of Multiple Instance Learning. It covers foundational concepts, traditional methods, recent advancements, and future directions. By providing a holistic view of MIL's dynamic landscape, this paper aims to inspire further innovation and exploration in this ever-evolving field. • A survey on the current state of the Multiple Instance Learning. • We provide applications of Multiple instance learning in various domains. • We discuss how existing supervised learning algorithms are modified to support MIL. • Publicly available datasets and open research challenges in MIL are discussed. [ABSTRACT FROM AUTHOR]

The second-leading cause of death for women is breast cancer. Consequently, a precise early diagnosis is essential. With the rapid development of artificial intelligence, computer-aided diagnosis can efficiently assist radiologists in diagnosing breast problems. Mammography images, breast thermal images, and breast ultrasound images are the three ways to diagnose breast cancer. The paper will discuss some recent developments in machine learning and deep learning in three different breast cancer diagnosis methods. The three components of conventional machine learning methods are image preprocessing, segmentation, feature extraction, and image classification. Deep learning includes convolutional neural networks, transfer learning, and other methods. Additionally, the benefits and drawbacks of different methods are thoroughly contrasted. Finally, we also provide a summary of the challenges and potential futures for breast cancer diagnosis. [ABSTRACT FROM AUTHOR]

In a constant evolving world, change detection is of prime importance to keep updated maps. To better sense areas with complex geometry (urban areas in particular), considering 3D data appears to be an interesting alternative to classical 2D images. In this context, 3D point clouds (PCs), whether obtained through LiDAR or photogrammetric techniques, provide valuable information. While recent studies showed the considerable benefit of using deep learning-based methods to detect and characterize changes into raw 3D PCs, these studies rely on large annotated training data to obtain accurate results. The collection of these annotations are tricky and time-consuming. The availability of unsupervised or weakly supervised approaches is then of prime interest. In this paper, we propose an unsupervised method, called DeepCluster 3D Change Detection (DC3DCD), to detect and categorize multiclass changes at point level. We classify our approach in the unsupervised family given the fact that we extract in a completely unsupervised way a number of clusters associated with potential changes. Let us precise that in the end of the process, the user has only to assign a label to each of these clusters to derive the final change map. Our method builds upon the DeepCluster approach, originally designed for image classification, to handle complex raw 3D PCs and perform change segmentation task. An assessment of the method on both simulated and real public dataset is provided. The proposed method allows to outperform fully-supervised traditional machine learning algorithm and to be competitive with fully-supervised deep learning networks applied on rasterization of 3D PCs with a mean of IoU over classes of change of 57.06% and 66.69% for the simulated and the real datasets, respectively. The code is available at https://github.com/idegelis/torch-points3d-dc3dcd. • DC3DCD, unsupervised learning method for multiclass 3D point cloud change detection. • Revisiting the DeepCluster principle for change segmentation in raw 3D point clouds. • Results are better than classical supervised machine learning methods. • Results are equal to rasterization-based deep learning supervised networks. • In a binary-to-multiclass scenario, DC3DCD competes with fully supervised methods. [ABSTRACT FROM AUTHOR]

Neurons are the fundamental units of neural networks. In this paper, we propose a method for explaining neural networks by visualizing the learning process of neurons. For a trained neural network, the proposed method obtains the features learned by each neuron and displays the features in a human-understandable form. The features learned by different neurons are combined to analyze the working mechanism of different neural network models. The method is applicable to neural networks without requiring any changes to the architectures of the models. In this study, we apply the proposed method to both Fully Connected Networks (FCNs) and Convolutional Neural Networks (CNNs) trained using the backpropagation learning algorithm. We conduct experiments on models for image classification tasks to demonstrate the effectiveness of the method. Through these experiments, we gain insights into the working mechanisms of various neural network architectures and evaluate neural network interpretability from diverse perspectives. [ABSTRACT FROM AUTHOR]

Roundwood sorting is still a manual process in many Swiss sawmills, requiring employees to visually inspect and categorize thousands of logs per day. The heavy workload can be both physically and mentally taxing and can lead to increased rates of human error. State-of-the-art automation systems like X-ray log scanners are expensive and difficult to integrate into existing process lines. This paper proposes a novel recommendation system that leverages recent advances in image classification to automate roundwood classification by quality and species. The system integrates a camera to capture cross-sectional images of logs and record numerical data, such as length, taper, and diameter. The analysis of the resulting dataset highlights the challenges of data imbalance and noise, which makes classification difficult and, in some cases, impossible. However, by using selected datasets with reduced noise, state-of-the-art Convolutional Neural Networks (CNNs) can extract quality and species features. Quality models learn from a manually selected and simplified dataset, featuring samples that experts can clearly classify based on the image's information. Species models are trained on a label-noise-reduced dataset, reflecting real-world complexity. The accuracy on the selected dataset for three quality classes is 80%. The species determination is less challenging and reaches 91% accuracy on a synchronized dataset for the main species spruce and fir. Overall, this paper highlights the potential of Machine Learning in augmenting the roundwood sorting processes and presents a novel system that can improve the efficiency and accuracy of the process. [Display omitted] • Automation of roundwood sorting: Replaces manual sorting with image-based AI. • Integrated camera system in roundwood sorting to collect labeled dataset. • Species prediction: 91% accuracy in spruce–fir distinction. • Quality prediction on complexity reduced dataset: 80% accuracy between three main quality levels. • Efficient, adaptable & scalable system which is easy to integrate into existing process lines. [ABSTRACT FROM AUTHOR]

In this paper, we use artificial intelligence image recognition to obtain Digestive endoscopy image, and process the image based on 5G Deep learning edge algorithm to judge the disease type of the patient, and then consider the treatment plan. The combination of body area network and edge computing technology can meet the demand of low delay in body area network. In this case, the resource constrained body area network gateway node can process the physiological data collected by it into an offloadable task, and then unload the task and data to the edge computing node according to a certain strategy. The edge computing node completes the corresponding task processing and data storage, and finally provides the results to the relevant medical institutions and body area network users for reading auxiliary diagnosis and treatment of diseases. Studies have shown that 25% of patients with colon polyps have CD4 cells in peripheral blood based on 5G deep learning edge algorithm under artificial intelligence image recognition of Digestive endoscopy. The number of lymphocyte in group of differentiation was less than 200/μL, and the blood RNA in 92.3% patients was lower than 100 IU/ml, while fam CTP (A-cyclic peptide) was lower than 100 IU/ml. Opportunistic infections of the intestine and viruses can directly cause enteropathy because the fluorescence intensity of the probe is essentially unchanged and cannot form a triple helix structure. In terms of feature recognition accuracy, the 5G deep learning edge algorithm in this paper improves accuracy by 68% compared to the simple Yolo algorithm, and is similar in speed. Compared with RCNN algorithm, the accuracy and speed are improved by 21% and 85% respectively. Therefore, the 5G deep learning edge algorithm based on artificial intelligence image recognition has the advantages of accuracy and speed in digestive endoscopy of intelligent medical. [ABSTRACT FROM AUTHOR]

To enhance the accuracy of object recognition, various combination of recognition algorithms are used in recent literature. In this paper coherence of Discrete Wavelet Transform (DWT) is combined with four different algorithms: error vector of principal component analysis (PCA), eigen vector of PCA, eigen vector of Linear Discriminant Analysis (LDA) and Convolutional Neural Network (CNN) then combination of four results are done using entropy of detection probability and Fuzzy system. From this research the accuracy of recognition is found dependent on image and diversity of database. The combined method of the paper provides recognition rate of 89.56% for the worst case and 93.34% for the best case both can be said better in comparison with the previous works where individual method has been implemented on a specific set of images. [ABSTRACT FROM AUTHOR]

Lifelong learning represents an emerging machine learning paradigm that aims at designing new methods providing accurate analyses in complex and dynamic real-world environments. Although a significant amount of research has been conducted in image classification and reinforcement learning, very limited work has been done to solve lifelong anomaly detection problems. In this context, a successful method has to detect anomalies while adapting to changing environments and preserving knowledge to avoid catastrophic forgetting. While state-of-the-art online anomaly detection methods are able to detect anomalies and adapt to a changing environment, they are not designed to preserve past knowledge. On the other hand, while lifelong learning methods are focused on adapting to changing environments and preserving knowledge, they are not tailored for detecting anomalies, and often require task labels or task boundaries which are not available in task-agnostic lifelong anomaly detection scenarios. This paper proposes VLAD, a novel V AE-based L ifelong A nomaly D etection method addressing all these challenges simultaneously in complex task-agnostic scenarios. VLAD leverages the combination of lifelong change point detection and an effective model update strategy supported by experience replay with a hierarchical memory maintained by means of consolidation and summarization. An extensive quantitative evaluation showcases the merit of the proposed method in a variety of applied settings. VLAD outperforms state-of-the-art methods for anomaly detection, presenting increased robustness and performance in complex lifelong settings. • VAE-based Lifelong Anomaly Detection (VLAD) for complex task-agnostic scenarios. • Model update strategy based on lifelong change point detection and experience replay. • A hierarchical memory maintained by means of consolidation and summarization. • Extensive evaluation showing VLAD's strengths in applied real-world evolving domains. • VLAD yields a robust task-agnostic lifelong anomaly detection performance. [ABSTRACT FROM AUTHOR]

Deep learning (DL) has been shown to be very effective for many application domains of machine learning (ML), including image classification, voice recognition, natural language processing, and bioinformatics. The success of DL techniques is directly related to the availability of large amounts of training data. However, in many cases, the data are sensitive to the users and should be protected to preserve the privacy. Privacy-preserving deep learning (PPDL) has thus become a very active research field to ensure the training process and use of DL models are productive without exposing or leaking information about the data. This paper aims to provide a comprehensive survey of PPDL. We concentrate on the risks that affect data privacy in DL and conduct a detailed investigation into the models that ensure privacy. Finally, we propose a set of evaluation criteria, detailing the advantages and disadvantages of the solutions. Based on the analyzed strengths and weaknesses, the paper has highlighted some important research problems and application cases that have not been studied and these point to certain open research directions. [ABSTRACT FROM AUTHOR]

• This paper provides an in-depth review of the latest deep learning methods for use in big data analytics. • Explain the importance of deep learning, its taxonomy, and big data analytics techniques. • Explores deep learning approaches in IoT data applications, including their complexities and limitations. • Suggests using deep learning techniques in many data-intensive applications and benchmarked frameworks and datasets. • A comparison of established approaches with deep learning methods in big data analytics is offered. Deep learning (DL), as one of the most active machine learning research fields, has achieved great success in numerous scientific and technological disciplines, including speech recognition, image classification, language processing, big data analytics, and many more. Big data analytics (BDA), where raw data is often unlabeled or uncategorized, can greatly benefit from DL because of its ability to analyze and learn from enormous amounts of unstructured data. This survey paper tackles a comprehensive overview of state-of-the-art DL techniques applied in BDA. The main target of this survey is intended to illustrate the significance of DL and its taxonomy and detail the basic techniques used in BDA. It also explains the DL techniques used in big IoT data applications as well as their various complexities and challenges. The survey presents various real-world data-intensive applications where DL techniques can be applied. In particular, it concentrates on the DL techniques in accordance with the BDA type for each application domain. Additionally, the survey examines DL benchmarked frameworks used in BDA and reviews the available benchmarked datasets, besides analyzing the strengths and limitations of each DL technique and their suitable applications. Further, a comparative analysis is also presented by comparing existing approaches to the DL methods used in BDA. Finally, the challenges of DL modeling and future directions are discussed. [ABSTRACT FROM AUTHOR]

Automatic detection of Diabetic Retinopathy (DR) is critically important, as it is the primary reason of irreversible loss of vision in the economically active populations in the developed countries. Early detection of the onset of Diabetic Retinopathy can greatly benefit clinical treatment; although several different feature extraction methods have been proposed, the task of retinal image classification remains tedious even for trained clinicians. This paper emphasizes on Diabetic Retinopathy detection as well as the analysis of the different stages of DR, performed on fundus images using Deep Learning algorithms. Fundus images of the patient were provided as input to the developed model evaluated using the real-time dataset of the hospital. The proposed ResNet-18 architecture with swish function has achieved an accuracy of 93.51%, sensitivity of 93.42%, precision of 93.77% and F1-score of 93.59%. The paper concludes with a comparative study of Simple CNN, VGGNet-16, MobileNet-V2 and ResNet architectures and other state-of-art approaches, which highlights ResNet-18 with Swish as the most effective deep learning classifier model for DR detection. • Deep learning outperforms traditional screening for Diabetic Retinopathy, reducing ophthalmologists' workload. • ResNet18's residual connections mitigate the vanishing gradient problem and enable effective retinal image pattern learning. • Swish activation in ResNet18 enhances Diabetic Retinopathy detection with improved accuracy and faster training convergence. [ABSTRACT FROM AUTHOR]

Recently, there has been a substantial interest in on-device Machine Learning (ML) models to provide intelligence for the Internet of Things (IoT) applications such as image classification, human activity recognition, and anomaly detection. Traditionally, ML models are deployed in the cloud or centralized servers to take advantage of their abundant computational resources. However, sharing data with the cloud and third parties degrades privacy and may cause propagation delay in the network due to a large amount of transmitted data impacting the performance of real-time applications. To this end, deploying ML models on-device (i.e., on IoT devices), in which data does not need to be transmitted, becomes imperative. However, deploying and running ML models on already resource-constrained IoT devices is challenging and requires intense energy consumption. Numerous works have been proposed in the literature to address this issue. Although there are considerable works that discuss energy-aware ML approaches for on-device implementation, there remains a gap in the literature on a comprehensive review of this subject. In this paper, we provide a review of existing studies focusing on-device ML models for IoT applications in terms of energy consumption. One of the key contributions of this study is to introduce a taxonomy to define approaches for employing energy-aware on-device ML models on IoT devices in the literature. Based on our review in this paper, our key findings are provided and the open issues that can be investigated further by other researchers are discussed. We believe that this study will be a reference for practitioners and researchers who want to employ energy-aware on-device ML models for IoT applications. [ABSTRACT FROM AUTHOR]

Domain adaptation is a practicable tool in real world application where there exists data scarcity. Multi-source domain adaptation attracts increasing attention due to its ability to enrich transfer knowledge by combining information from multiple domains. However, knowledge transfer can trigger privacy concerns by accessing source data. In addition, multiple domains can have label heterogeneity problem. In this paper, to solve the mentioned problems, we conduct an incomplete multi-source domain adaptation (IMSDA) method which can address transfer learning with and without the access to source data. As far as we are aware, this is the first work handling source-free incomplete domain adaptation. To take the benefits of multiple sources, multi-task learning is adopted to learn a general source model which can perform on multiple domains. A data matching strategy with and without source data forcing target sample to source latent feature space is developed to combine with self-supervision to adapt source model to the target domain. Experiments on real-world datasets indicate the superiority of the proposed method. [ABSTRACT FROM AUTHOR]

In recent years, deep learning (DL) methods have proven their efficiency for various computer vision (CV) tasks such as image classification, natural language processing, and object detection. However, training a DL model is expensive in terms of both complexities of the network structure and the amount of labeled data needed. In addition, the imbalance among available labeled data for different classes of interest may also adversely affect the model accuracy. This paper addresses these issues using a new convolutional neural network (CNN) based architecture. The proposed network incorporates both spatial and spectral information that combines two sub-networks: spatial-CNN and spectral-CNN. The spectral-CNN extracts spectral information, while spatial-CNN captures spatial information. Moreover, to make the features more robust, a multiscale spatial CNN architecture is introduced using different kernels. The final feature vector is formed by concatenating the outputs obtained from both spatial-CNN and spectral-CNN. To address the data imbalance problem, a generative adversarial network (GAN) was used to generate data for the underrepresented class. Finally, relatively a shallower network architecture was used to reduce the number of parameters in the network and improve the processing speed. The proposed model was trained and tested on Senitel-2 images for the classification of the debris-covered glacier. The results showed that the proposed method is well-suited for mapping and monitoring debris-covered glaciers at a large scale with high classification accuracy. In addition, we compared the proposed method with conventional machine learning approaches, support vector machine (SVM), random forest (RF) and multilayer perceptron (MLP). [ABSTRACT FROM AUTHOR]

Simultaneous localization and mapping algorithm is a technique to simultaneously construct a map of the environment by the sensors it carries. In this paper, we focus on the SLAM (Simultaneous Localization and Mapping) algorithm based on machine vision-depth information, overlap region determination among local sub-maps, and fusion algorithm of 3D dense maps, and analyze four modules of front-end visual odometry, back-end optimization, loopback detection, and map construction. The front-end visual odometry adopts the ORB (Oriented FAST and Rotated BRIEF) algorithm based on the feature point method to extract and match the features between image frames and uses the iterative nearest point method to solve the camera motion between adjacent images. Meanwhile, a 3D feature descriptor based on the histogram of spatial distribution is designed to encode the feature information around the key points, and the Kd-Tree algorithm is used for radial nearest neighbor search to complete the feature matching. Finally, the 3D map is fused by the 3D-ICP algorithm. The experiments prove that the low-dimensional local feature descriptor designed has good descriptiveness and is superior, and the 3D map fusion based on this descriptor has a good fusion effect and can meet the accuracy requirements of practical applications. [ABSTRACT FROM AUTHOR]

Image classification, which is a fundamental element of Computer Vision (CV) and Artificial Intelligence (AI), has been researched intensively in numerous domains and embedded in many products. However, with the exponential increase in the number of images and the complexity of the required tasks, deep-learning classification algorithms demand more intensive resources and computational power to train the models and update the massive number of parameters. Quantum computing is a new research technology with a promising capability of exponential speed up and operational parallelization with its unique phenomena including superposition and entanglement. Researchers have already started utilizing Quantum Deep Learning (QDL) and Quantum Machine Learning (QML) in image classification. Yet, to our knowledge, there exists no comprehensive published literature review on quantum image classification. Therefore, this paper analyzes the advances in this field by dividing the studies based on a unique taxonomy, discussing the limitations, summarizing essential aspects of each research, and finally, emphasizing the gaps, challenges, and recommendations. One of the key challenges presented in the paper is that quantum computers are in the Noisy Intermediate-Scale Quantum (NISQ) era, where they contain a limited number of noisy qubits, therefore challenging complex quantum classifiers and complex images from advanced datasets. This research recommends constructing a novel quantum image encoding method that adapts to the available number of qubits and enables RGB images as a critical contribution to the existing research. [ABSTRACT FROM AUTHOR]

[Display omitted] • Machine learning-based image recognition expedites the discovery and development of novel TMS catalysts for various applications. • The study finds weak C 4 H 4 S–M bonds on first-row TMSs, but identifies V 5 S 4 and Ti 8 Cu 3 S 16 as promising catalysts by exploring bridge sites for thiophene adsorption. • The Turi Create-based machine learning approach requires minimal training data and expertise, making it accessible to a broader range of researchers. • Machine-based identification of COHP plots for catalysts can be extended to the development of other catalyst types. This paper investigates the potential of transition metal sulfides (TMSs) as hydrodesulfurization (HDS) catalysts through electronic structure and bonding analysis. The HDS mechanism focuses on breaking and regenerating metal–sulfur (M−S) bonds at sulfur vacancies, known as active sites adhering to the Sabatier principle. Machine learning was employed to identify TMSs with crystal orbital Hamilton populations (COHP) similar to known catalysts like OsS 2 or Co–Mo–S, facilitating novel catalyst development. However, the weak thiophene adsorption on the top sites of the first-row TMS surfaces limits their catalytic potential. To address this limitation, adsorption on bridge sites was explored, leading to the discovery of V 5 S 4 and Ti 8 Cu 3 S 16 as highly promising HDS catalysts. These findings suggest that first-row TMSs can serve as cost-effective and efficient HDS catalysts, offering valuable insights for the development of other catalysts. [ABSTRACT FROM AUTHOR]

The aim of data science is to catch up with the data-intensive life style as well as the demand for decision support, which becomes common in various domains such as medical, education and other smart solutions. As such, high quality of data analysis is greatly desired for accurate and effective downstreaming exploitations. This is also true for the domain of astronomical survey like GOTO (Gravitational-wave Optical Transient Observer), where large amount of raw data has been collected daily. This is one of recognised projects that search for transient events with the new breed of optical survey telescopes that can detect the sky faster and deeper. This is accomplished by comparing the night-specific data with the reference such that new bright sources are obtained for further study. However, the huge size of data makes it difficult to sift by naked eyes, thus requiring an automated system. Yet, many conventional machine-learning models have been sub-optimal for this task, as true positives can hardly be recognised due to the nature of imbalance data. This motivates the exploration of convolutional neural networks or CNN for this binary classification problem. Based on existing technologies, the paper reports the original application of basic CNN model to a representative data, which has been designed and generated within the GOTO project. In addition to the improvement over those previous works, this empirical study also includes details of parameter analysis, which will be useful for practice and further investigation. [ABSTRACT FROM AUTHOR]

• Class-incremental learning frameworks mitigate the catastrophic forgetting problems of continual plant diagnosis. • Adversarial training significantly improves the generalization of class-incremental classification. • The advantage of adversarial training based class incremental learning is more pronounced with less exemplar storage. • Employing the adversarial training after the conventional training to finetune the parameters is efficient. Early diagnosis of plant disease is of utmost importance to improve the crop yield and the crop quality. In the early stage of disease infestation, characteristics of the plant leaves could be used to indicate the disease category. As a reliable tool for leaf-image classification, deep learning approaches have achieved a wide attention recently. These data-driven approaches could extract the characteristics by the mathematical model from a diseased leaf-image dataset automatically, then output the prediction result which belongs to the dataset label. However, if the model learns a new disease incrementally from future data, it might forget previously learned knowledge. In this paper, we investigate the rehearsal-based class-incremental learning approaches for plant disease classification, in order to mitigate the catastrophic forgetting while learning the knowledge of a new disease. Moreover, we explore the effectiveness of class-incremental learning approaches on images with significantly different distribution. To improve the prediction and generalization performance of these out-of-distribution leaf images captured in complex scenes of the real-world, we propose a new rehearsal-based class-incremental learning approach, which is termed as adversarial class-incremental learning. The approach is based on a proposed adversarial training algorithm, which finetunes the trained parameters by projected gradient descent instead of learning the weights directly by adversarial samples. In the experiment of a class-incremental learning task with 30 categories, the proposed approach achieved a predictive accuracy of 86.8% on the leaves of PlantVillage dataset, and an accuracy of 66.6% on the leaf-image dataset with cluttered backgrounds, outperforming the existing suboptimal method which achieved the accuracies of 83.3% and 57.6%. The satisfactory performance of our proposed method provides an opportunity of continual diagnosis for agricultural application. [ABSTRACT FROM AUTHOR]

Compared to white-box adversarial attacks, black-box adversarial attacks are more applicable in practical scenarios and have received significant attention. However, most existing black-box attacks are optimized at the output layer, and the generated adversarial samples are difficult to transfer from the surrogate model to the target model. Existing adversarial attacks indiscriminately perturb features in intermediate layers, which are prone to fall into local optima of surrogate models and have limited transferability. In this paper, we propose a transferable targeted adversarial attack framework based on a feature-aware triplet, which achieves a better tradeoff between attack ability and transferability by perturbing salient features and constructing representative sample features. The importance-weighted optimization objective interferes with the salient object-aware features of images in a targeted manner and guides adversarial perturbations in the optimization process to pull the intermediate features of samples toward a target class while pushing them far from a source class to achieve targeted transferable attacks. Moreover, a construction method of a feature library with the weighted average of feature importance is built to obtain more expressive intermediate features of the target and source classes. The modified target features and source features are fed into the triplets to guide the optimization objectives to find more transferable adversarial samples. Extensive experiments on the ImageNet-compatible dataset verify the effectiveness of the proposed method, e.g., improving the untargeted success rate by 1.8% and the targeted success rate by 2.3% against normally trained models as compared to the existing methods. • An attack framework based on a feature-aware triplet is proposed. • Achieve a better tradeoff between attack capability and transferability. • A weighted average method for constructing a feature library is designed. • Define more expressive sample spaces. • Extended experiments exhibit superior transferability of the proposed FTA. [ABSTRACT FROM AUTHOR]

• Almost all these works above take the independent and identically distributed (i.i.d.) assumption, i.e., regarding all samples in various computer vision tasks as i.i.d., but they are not i.i.d. samples and the samples of many machine learning applications such as speech recognition, system diagnosis, and market forecasting are also proved to be non-i.i.d. samples. In particular, with the advent of the big data era, multi-class learning problems-such as text or image classification - can involve tens or hundreds of thousands of classes, and the samples are usually spatially-connected or temporally-correlated with their physically-connected neighbours, which can not be directly handled by these existing methods or algorithms. This implies that many samples are not satisfy the i.i.d. assumption or this assumption is very restrictive and can not be strictly justified in real-world problems. In addition, the AIO methods above are theoretically reasonable (i.e., Fisher consistent) and experiments indicate that they can lead to well-generalizing hypothesis, but efficient algorithms are lacking so far. Then three problems are posed: • Whether the AIO-MSVM algorithm based on non-i.i.d. samples is consistent? If this algorithm is consistent, what is the learning rate of this algorithm with non-i.i.d. samples? Whether the efficient AIO-MSVM algorithm can be proposed (or whether the AIO-MSVM algorithm with i.i.d. samples can be improved)? • The goal of this paper is to answer these problems above. Now, let' s highlight the main features of this paper. • The generalization bound of AIO-MSVM algorithm with u.e.M.c. samples is established and the fast learning rate is obtained. The AIO-MSVM algorithm with u.e.M.c. samples is proved to be consistent. To our knowledge, these studies are the first works on this topic. • A new algorithm, called AIO-MSVM algorithm based on q -times Markovian resampling (AIO-Mar- q), is proposed. The effectiveness and efficiency of the proposed algorithm are validated by experiments on public datasets. The generalization performance of "All-in-one" Multi-class SVM (AIO-MSVM) based on uniformly ergodic Markovian chain (u.e.M.c.) samples is considered. We establish the fast learning rate of AIO-MSVM algorithm with u.e.M.c. samples and prove that AIO-MSVM algorithm with u.e.M.c. samples is consistent. We also propose a novel AIO-MSVM algorithm based on q -times Markovian resampling (AIO-MSVM-MR), and show the numerical investigation on the learning performance of AIO-MSVM-MR based on public datasets. The experimental studies indicate that compared to the classical AIO-MSVM algorithm and other MSVM algorithms, the proposed AIO-MSVM-MR algorithm has not only smaller misclassification rate, but also less sampling and training total time. We present some discussions on the case of unbalanced training samples, the choices of q and two technical parameters, and present some explanations on the learning performance of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Multi-label image recognition is a basic and challenging task in computer vision and multimedia fields. Graph Convolutional Networks (GCNs) are often used to learn the multi-label semantic features and multi-label dependency. Although the label semantic features in GCNs can learn the global image visual representation well, they are rarely used on the local image regions. Therefore, we try to use GCNs to learn global and local features at the same time, and make a balance between them. In this paper, we give a multi-scale semantic attention model MS-SGA-GCN including three main modules (i.e., MS, SGA and GCN) for multi-label image recognition. The Multi-Scale module (MS) utilizes feature maps of different sizes to obtain global features and have strong generalization capabilities. Semantic Guide Attention module (SGA) applies the label embeddings learned by GCNs to guide the generation of the cross-modality class-specific attention maps, which can discover the locations of semantically related regions for each label. Experiments show that our model on two datasets MS-COCO and PASCAL VOC2007 separately achieves the classification accuracy by 83.4 % and 94.2 % , which has a competitive advantage over other mainstream models. [ABSTRACT FROM AUTHOR]

Deep representation learning is a subfield of machine learning that focuses on learning meaningful and useful representations of data through deep neural networks. However, existing methods for semantic classification typically employ pre-defined target codes such as the one-hot and the Hadamard codes, which can either fail or be less flexible to model inter-class correlation. In light of this, this paper introduces a novel learnable target coding as an auxiliary regularization of deep representation learning, which can not only incorporate latent dependency across classes but also impose geometric properties of target codes into representation space. Specifically, a margin-based triplet loss and a correlation consistency loss on the proposed target codes are designed to encourage more discriminative representations owing to enlarging between-class margins in representation space and favoring equal semantic correlation of learnable target codes respectively. Experimental results on several popular visual classification and retrieval benchmarks can demonstrate the effectiveness of our method on improving representation learning, especially for imbalanced data. • We propose an auxiliary feature regularization for deep representation learning. • We propose a learnable target coding scheme for capturing target correlations. • Our method consistently outperforms state-of-the-art on popular benchmarks. [ABSTRACT FROM AUTHOR]

Addressing the limitations of most few-shot learning (FSL) methods, particularly their insufficient single-feature discriminability and generalization during pre-training encoding, this paper introduces a novel approach: Two-Stage Training based Metric Fusion Learning (TST_MFL) for few-shot image classification. TST_MFL innovatively integrates a two-stage fusion metric learning strategy, enhancing both pre-training and meta-training phases to overcome the prevalent focus on single-level feature extraction in current few-shot image classification methods. In pre-training stage, we employ prediction distillation loss and local difference loss to facilitate mutual learning between global and local features, thus obtaining more discriminative features. Moreover, meta-training further enhances feature generalization and integrates global–local metrics to improve few-shot classification accuracy. Our approach demonstrates competitive performance across three fine-grained datasets and one coarse-grained dataset. Notably, experimental results in 1-shot and 5-shot settings of fine-grained datasets underscore the superiority of TST_MFL. The code for this work is available at: https://github.com/ZitZhengWang/TST_MFL. • The TST_MFL proposed pioneers a targeted fusion metric strategy in both pre-training and meta-training stages, effectively addressing the limitations of single feature representation in few-shot image classification. • The DPPTNet is introduced, integrating global–local information via prediction distillation loss and local difference loss, yielding proficient global and local feature simultaneously. • The GLMFNet in meta-training stage is proposed, enhancing similarity measurement accuracy through the fusion of global and local metrics. • Extensive experiments on four benchmark datasets validate the efficacy of our approach. [ABSTRACT FROM AUTHOR]

Choroidal neovascularization (CNV) is an eye disease that can cause vision loss. Automatic CNV classification in OCT images is crucial in the treatment of CNV. However, two problems arise for CNV classification in OCT images. The subtle visual differences between different CNV types render classification difficult. Additionally, it is difficult to obtain sufficient labeled data, which results in performance degradation. In order to solve these two problems, a discriminative atom-embedding relation dual network is proposed in this paper. Considering that semi-supervised learning (SSL) is an effective machine learning framework to make full use of limited labeled data and a large amount of unlabeled data, the proposed network is developed within an SSL framework. To capture the visual differences, novel discriminative atoms are first introduced to mine discriminative information between different CNV types. Subsequently, a relation module is incorporated to embed the learned discriminative atom information into the features. This makes the learned features capable of distinguishing between different CNV types. Moreover, a novel relation consistency loss is proposed to further improve the robustness of the learned features. Experimental results on private and public datasets demonstrate the effectiveness of the proposed method. • A novel discriminative atoms embedding relation dual network is proposed to address the challenge of inter-class similarities in CNV classification with limited labeled data. • To represent the discriminative details of CNV latently, discriminative atoms are introduced with incorporation of dynamic discriminative learning mechanism and constraint of discrimination loss. • In order to learn the discriminative features necessary to distinguish different CNV class, a novel relation learning module is introduced to embed the discriminative atoms into the learned features. • A dual mean-teacher network architecture is developed, and a relation consistency loss is proposed to guarantee that discriminative features can be learned during the utilization of unlabeled data. [ABSTRACT FROM AUTHOR]

Zero-shot learning aims to recognize objects from novel concepts through the model trained on seen class data and assisted by the semantic descriptions. Though it breaks the serious reliance on training data, it still fails to deal with the sequential streaming data in the open world. In this paper, we focus on the incremental zero-shot learning (IZSL) where the seen data arrives in the form of task sequence. In each incremental task, we only have access to the seen data of current task. The IZSL methods aim to depict the characteristics of current seen classes and avoid forgetting the previous ones, in the meantime, learn to generalize to the unseen classes. We summarize the challenges in IZSL as semantic drift which is further divided into task-recency bias and seen-class bias. To solve these issues, we propose a novel IZSL method termed as CRAA. Specifically, CRAA constructs consistent representations with satisfactory discrimination and generalization for all the seen and unseen classes. Based on these representations, CRAA learns a prototype classifier with a novel adaptive adjustment strategy to alleviate the task-recency bias and seen-class bias. Note that CRAA only needs a limited memory footprint to store the fixed scale model, and meets the demands of both memory restriction and data security in industry. We have conducted extensive experiments to evaluate our method on three widely used datasets. The results prove our method is superior to all the compared methods with significant improvements. Code is available at: https://github.com/changniu54/CRAA-Master. • We summarize the IZSL challenges as task-recency bias and seen-class bias. • We propose to learn consistent representation joint an adaptive adjustment strategy. • The method needs limited memory footprint to store the fixed scale model. • The method meets the demands of both memory restriction and data security. [ABSTRACT FROM AUTHOR]

The process of applying machine learning algorithms to practical problems can be a challenging and tedious task for non-experts. Previous research has sought to alleviate this burden by introducing automated machine learning techniques, including Network Architecture Search (NAS) and Differentiable Architecture Search (DARTS). However, these methods use a fixed number of layers and predefined skip connections which impose limitations on the generation of an optimal network architecture. In this paper, we propose a novel approach called Greedy Layer-wise Network Architecture Search (GLNAS), which trains network layers one after another and evaluates the network's performance after each layer is added. GLNAS also assesses the effectiveness of skip connections between layers by testing various outputs of previous layers as an input to the current layer. Our experiment results demonstrate that the network generated by GLNAS requires fewer parameters (i.e., 3.5 millions in both CIFAR-10 and CIFAR-100 datasets) and GPU resources during the searching phase (i.e., 0.17 and 0.24 GPU days in CIFAR-10 and CIFAR-100 datasets respectively) than many existing methods. • Our GLNAS shows the effectiveness of layer-wise training regardless of activations. • GLNAS offers flexibility of skip connections which achieves highly robust networks. • The model generated from GLNAS method spends low search cost and use less parameters. [ABSTRACT FROM AUTHOR]

In the paper, we present a new transfer subspace learning algorithm termed discriminative transfer regression (DTR) for cross-domain image recognition, in which low-rank representation (LRR), discriminative regression, local geometry preserving, and different norm minimization are integrated in to a united framework for transfer learning. Firstly, both the global structure and the local geometry information of the original data are preserved by imposing low-rank and sparse constraints on the reconstruction coefficient matrix. Secondly, DTR algorithm can overcome the disturbance of outliers and noises. Thirdly, the local manifold structure of the observed samples with the same semantics from the source and target domains is captured by the adaptive weight graph. Fourthly, the discriminative information of the samples from the source domain is encoded to the target domain based on ridge regression (RR). Meanwhile, the small-class problem confronted by RR and its extensions, that the obtained projection matrix is limited by the number of classes, can be effectively solved. In addition, the convergence and computational complexity of DTR algorithm is analyzed. An extensive range of experiments on several cross-domain image databases demonstrate the superiority of the DTR algorithm. [ABSTRACT FROM AUTHOR]

Active learning alleviates labeling costs by selecting and labeling the most informative examples from an unlabeled pool. However, most existing active learning approaches estimate informativeness with uncalibrated confidence, resulting in unreliable informativeness estimation. These approaches generally ignored two significant issues caused by uncalibrated confidence methods. Firstly, the average uncalibrated confidence generated by modern neural networks is usually higher than the accuracy. Secondly, examples located near the decision boundaries are unstable during prediction when the target model updates parameters in the last several epochs, even throughout the training process. This phenomenon, caused by the forgetting characteristic of neural networks, has a significant impact on some specific models that estimate the informativeness by predicted probability vectors or pseudo labels. To address these issues, in this paper, we propose a novel active learning approach to reliably estimate informativeness with calibrated confidence. Specifically, we integrate the intermediate predictions for each unlabeled example, generated by the target model during the training process, to generate calibrated confidence. The calibrated confidence can capture a tendentious label from an indecisive subset of the class space. We show that the calibrated confidence with tendentiousness can maintain the ability of correct predictions. The empirical results demonstrate that our approach outperforms the state-of-the-art active learning methods on image classification tasks. • The active learning algorithm selects unstable examples by calibrated confidence.. • The confidence calibration method only requires training a single model. • Extensive experiments show the competitive performance of the proposed approach. [ABSTRACT FROM AUTHOR]

This paper for the first time treats the interpretation of electrochemical noise time-frequency spectra as an image classification problem. It investigates the application of a convolutional neural network (CNN) for deep learning image classification of electrochemical noise time-frequency transient information. Representative slices of these spectra were selected by our transient analysis technique and served as input images for the CNN. Corrosion data from two types of pitting corrosion processes serve as test cases: AISI304 and AA2024-T3 immersed in a 0.01M HCl and 0.1M NaCl solution between 0 and 1ks after immersion, respectively. Continuous wavelet transform (CWT) spectra and modulus maxima (MM) are used to train the CNN, either individually or in a combined form. The classification accuracy of the CNN trained with the combined dataset is 0.97 and with the two individual datasets 0.72 (only CWT spectrum) and 0.84 (only MM). The ability to additionally classify a more progressed form of pitting corrosion of AA2024-T3 between 9 and 10ks after immersion indicates that the proposed method is sufficiently robust using combined datasets with CWT spectra and MM. The pitting processes can effectively be detected and classified by the proposed method. The most important contribution of the present work is to introduce a novel procedure that decreases the classical need for large amounts of raw data for training and validation purposes, while still achieving a satisfactory classification robustness. A relatively small number of individual signals thereby generates a multitude of input images that still contain all relevant kinetic information about the underlying chemo-physical process. [Display omitted] • Electrochemical noise interpretation is treated as an image classification problem. • Continuous wavelet transform spectra are divided into single transient spectra. • A small number of individual signals generates a multitude of input images. • Amount of raw data needed for training the convolutional neural network is reduced. • Modulus maxima increase the classification accuracy of pitting corrosion. [ABSTRACT FROM AUTHOR]

Gentrification is a complex and context-specific process that involves changes in the built environment and social fabric of neighborhoods, often resulting in the displacement of vulnerable communities. Machine Learning (ML) has emerged as a powerful predictive tool that is capable of circumventing the methodological challenges that historically held back researchers from producing reliable forecasts of gentrification. Additionally, computer vision ML algorithms for landscape character assessment, or deep mapping, can now capture a wider range of built metrics related to gentrification-induced redevelopment. These novel ML applications promise to rapidly progress our understandings of gentrification and our capacity to translate academic findings into more productive direction for communities and stakeholders, but with this sudden development comes a steep learning curve. The current paper aims to bridge this divide by providing an overview of recent progress and an actionable template of use that is accessible for researchers across a wide array of academic fields. As a secondary point of emphasis, the review goes over Explainable Artificial Intelligence (XAI) tools for gentrification models and opens up discussion on the nuanced challenges that arise when applying black-box models to human systems. • Bolstering deep mapping gentrification analyses with more interpretable algorithms such as those implemented for semantic image classification can increase model reliability and accuracy. • Explainable Artificial Intelligence (XAI) methods are an important addition to a black-box model, particularly for human-based systems like gentrification, but few studies incorporate explainability techniques. • Machine learning-based gentrification analyses benefit from context-specific, qualitative knowledge of the study area. [ABSTRACT FROM AUTHOR]

Machine learning (ML) and deep learning (DL) models have been extensively explored for the early diagnosis of various cancer diseases, including Leukemia, with many of them achieving significant performance improvements comparable to those of human experts. However, challenges like limited image data, inaccurate annotations, and prediction reliability still hinder their broad implementation to establish a trustworthy computer-aided diagnosis (CAD) system. This paper introduces a novel weighted-average ensemble model for classifying Acute Lymphoblastic Leukemia, along with a reliable Computer-Aided Diagnosis (CAD) system that combines the strengths of both ML and DL approaches. Initially, a variety of filtering methods are extensively analyzed to determine the most suitable image representation, with subsequent data augmentation techniques to expand the training data. Second, a modified VGG-19 model was proposed with fine-tuning that was utilized as a feature extractor to extract meaningful features from the training samples. Third, A small-loss approach and probabilistic local outlier factor (pLOF) have been developed on the extracted features to address the label noise issue. Fourth, we proposed an weighted-average ensemble model based on the top five models as base learners, with weights calculated based on their model uncertainty to ensure reliable predictions. Fifth, we calculated Shapley values based on cooperative game theory and performed feature selection with different feature combinations to determine the optimal number of features using SHAP. Finally, we integrate these strategies to develop an interpretable CAD system. This system not only predicts the disease but also generates Grad-CAM images to visualize potential affected areas, enhancing both clarity and diagnostic insight. All of our code is provided in the following repository: https://github.com/taareek/leukemia-classification [ABSTRACT FROM AUTHOR]

In 2016, Google proposed Federated Learning (FL) as a novel paradigm to train Machine Learning (ML) models across the participants of a federation while preserving data privacy. Since its birth, Centralized FL (CFL) has been the most used approach, where a central entity aggregates participants' models to create a global one. However, CFL presents limitations such as communication bottlenecks, single point of failure, and reliance on a central server. Decentralized Federated Learning (DFL) addresses these issues by enabling decentralized model aggregation and minimizing dependency on a central entity. Despite these advances, current platforms training DFL models struggle with key issues such as managing heterogeneous federation network topologies, adapting the FL process to virtualized or physical deployments, and using a limited number of metrics to evaluate different federation scenarios for efficient implementation. To overcome these challenges, this paper presents Fedstellar, a novel platform designed to train FL models in a decentralized, semi-decentralized, and centralized fashion across diverse federations of physical or virtualized devices. Fedstellar allows users to create federations by customizing parameters like the number and type of devices training FL models, the network topology connecting them, the machine and deep learning algorithms, or the datasets of each participant, among others. Additionally, it offers real-time monitoring of model and network performance. The Fedstellar implementation encompasses a web application with an interactive graphical interface, a controller for deploying federations of nodes using physical or virtual devices, and a core deployed on each device, which provides the logic needed to train, aggregate, and communicate in the network. The effectiveness of the platform has been demonstrated in two scenarios: a physical deployment involving single-board devices such as Raspberry Pis for detecting cyberattacks and a virtualized deployment comparing various FL approaches in a controlled environment using MNIST and CIFAR-10 datasets. In both scenarios, Fedstellar demonstrated consistent performance and adaptability, achieving F 1 s c o r e s of 91%, 98%, and 91.2% using DFL for detecting cyberattacks and classifying MNIST and CIFAR-10, respectively, reducing training time by 32% compared to centralized approaches. • Open-source platform for training FL models in a decentralized fashion. • Support for diverse FL architectures, network topologies, datasets, and ML/DL models. • User-friendly setup and efficient monitoring of FL processes. • Successful deployment across diverse physical and virtualized devices. • Promising performance in tasks like cyberattack detection and image classification. [ABSTRACT FROM AUTHOR]

This paper introduces a deep learning method for image classification that leverages knowledge formalized as a graph created from information represented by pairs attribute/value. The proposed method investigates a loss function that adaptively combines the classical cross-entropy commonly used in deep learning with a novel penalty function. The novel loss function is derived from the representation of nodes after embedding the knowledge graph and incorporates the proximity between class and image nodes. Its formulation enables the model to focus on identifying the boundary between the most challenging classes to distinguish. Experimental results on several image databases demonstrate improved performance compared to state-of-the-art methods, including classical deep learning algorithms and recent algorithms that incorporate knowledge represented by a graph. [ABSTRACT FROM AUTHOR]

In the medical domain, obtaining training images on a large scale is difficult due to privacy and cost issues, and as further disease incidence varies widely, class imbalance has been a critical problem. Importantly, securing samples for rare diseases will always be difficult, but a diagnosis model should be able to predict the diseased samples accurately. When a model is trained on imbalanced data, it is expected that (1) the model only exhibits biased results for the majority classes, and (2) the recognition performance for minority classes will decline rapidly, resulting in low specificity. We aim to improve the prediction performance of such few sample classes using our auxiliary data from an efficient generation model and style knowledge distillation. In this paper, we introduce a style knowledge distillation method to address the class imbalance problem. The proposed method generates auxiliary images in which the style of a training sample is transferred. We train our model with balanced training samples by distilling the styles of auxiliary images. We demonstrate that it considerably increases the accuracy of classes with few samples, notably without adversely impacting classes with large samples, using the APTOS2019, and ISIC2018. This approach is straightforward but very effective for medical image datasets with skewed class distributions. • We propose the auxiliary images based training algorithm to tackle the class imbalance problem in medical image classification. • We introduce a distillation algorithm to effectively learn auxiliary images. • We demonstrate that our proposed method performs well on class imbalanced medical image classification datasets. [ABSTRACT FROM AUTHOR]

In the paper, the automated evolutionary approach FEDOT-NAS for the design of convolutional neural networks is proposed. It allows building object recognition models for remote sensing tasks. The comparison of the proposed approach with state-of-the-art tools for neural architecture search is conducted for several examples of satellite-related datasets. The results of the experiments confirm the correctness and effectiveness of the proposed approach. The implementation of FEDOT-NAS is available as an open-source tool. [ABSTRACT FROM AUTHOR]

Image data collected after natural disasters play an important role in the forensics of structure failures. However, curating and managing large amounts of post-disaster imagery data is challenging. In most cases, data users still have to spend much effort to find and sort images from the massive amounts of images archived for past decades in order to study specific types of disasters. This paper proposes a new machine learning based approach for automating the labeling and classification of large volumes of post-natural disaster image data to address this issue. More specifically, the proposed method couples pre-trained computer vision models and a natural language processing model with an ontology tailed to natural disasters to facilitate the search and query of specific types of image data. The resulting process returns each image with five primary labels and similarity scores, representing its content based on the developed word-embedding model. Validation and accuracy assessment of the proposed methodology was conducted with ground-level residential building panoramic images from Hurricane Harvey. The computed primary labels showed a minimum average difference of 13.32% when compared to manually assigned labels. This versatile and adaptable solution offers a practical and valuable solution for automating image labeling and classification tasks, with the potential to be applied to various image classifications and used in different fields and industries. The flexibility of the method means that it can be updated and improved to meet the evolving needs of various domains, making it a valuable asset for future research and development. [ABSTRACT FROM AUTHOR]

[Display omitted] • A model between particle image grayscale features and LOI was investigated. • GA-ELM has higher accuracy, ability, and stability in detecting LOI than ELM. • A new method for LOI detection based on image recognition was investigated. Loss on ignition (LOI) is an essential evaluation index for fly ash triboelectric separation to recover unburned carbon products, and an important basis to regulate the technology parameters of triboelectric separation. Due to the problems of the traditional cauterization method, such as complex process, time-consuming, and error interference, the purpose of the paper is to introduce a new method for rapid detection of fly ash LOI based on image recognition and machine learning. An image recognition technique was applied to extract image grayscale features of unburned carbon particles recycled by triboelectric separation. The extreme learning machine (ELM) and genetic algorithm-optimized extreme learning machine (GA-ELM) neural networks established the relationship between grayscale feature parameters and LOI. Fly ash original samples and electrowinning experimental samples were used to evaluate the feasibility of 2 neural networks. The results showed that the evaluation indexes of GA-ELM network consisted of R2 is 95.218%, MAE is 0.0041546%, RMSE is 0.006709% and MAPE is 0.04213% at different voltages in the triboelectric separation experiments. The detection accuracy and detection ability of GA-ELM network was higher than ELM model. Therefore, the method can quickly and accurately detect the LOI of unburned carbon recovered from fly ash triboelectric separation. [ABSTRACT FROM AUTHOR]

Machine learning applied to computer vision and signal processing is achieving results comparable to the human brain due to the great improvements brought by deep neural networks (DNN). The majority of state-of-the-art architectures are DNN-related, but only a few explicitly explore the frequency domain to extract useful information and improve the results. This paper presents a new approach for exploring the Fourier transform of the input images, which is composed of trainable frequency filters that boost discriminative components in the spectrum. Additionally, we propose a cropping procedure to allow the network to learn both global and local spectral features of the image blocks. The proposed method proved to be competitive concerning well-known DNN architectures in the selected experiments, which involved texture classification, cataract detection, and retina image analysis, where there is a noticeable appeal for the frequency domain, with the advantage of being a lightweight model. • A new architecture for neural networks exploring the frequency domain is proposed. • Trainable frequency filters retrieve image discriminative features. • A block division scheme allows extracting local and global spectral features. • A frequency pooling technique reduces the model parameters and training time. • The proposed model reaches competitive results when compared to modern ConvNets. [ABSTRACT FROM AUTHOR]

Traditional machine learning methods require the assumption that training and test data are drawn from the same distribution, which proves challenging in real-world applications. Moreover, deep learning models require a substantial amount of labeled data for training in classification tasks and limited samples may lead to overfitting. In many real-world scenarios, there is an insufficient supply of labeled samples within the target domain for learning. Transfer learning offers an effective solution, allowing knowledge from a source domain to be transferred to a target domain. Additionally, data augmentation enhances model generalization by increasing data samples, particularly beneficial when dealing with limited target domain data. In this paper, we synergistically enhance the model's performance on classification tasks by integrating transfer learning techniques with a data augmentation strategy. By conducting numerous experiments across various datasets, we verified the effectiveness of our proposed approach. [ABSTRACT FROM AUTHOR]

Gradient-based optimization and control frameworks have been utilized in several applications. The learning rate parameter is typically chosen following a schedule or using methods such as line search to enhance the convergence rate. Recently, the machine learning community has developed methodologies for automated tuning of the learning rate, known as adaptive gradient methods. This paper develops a control theory-inspired framework for modeling adaptive gradient methods that solve non-convex optimization problems. We first model the adaptive gradient methods in a state–space framework, which allows us to present simpler convergence proofs of prominent adaptive optimizers, such as AdaGrad, Adam, and AdaBelief. The proposed framework is constructive because it allows synthesizing new adaptive optimizers. To illustrate this fact, we then utilize the transfer function paradigm from classical control to propose a new variant of Adam, coined AdamSSM, and prove its convergence. We add an appropriate pole-zero pair in the transfer function from squared gradients to the second moment estimate. Applications on benchmark machine learning tasks of image classification using CNN architectures and language modeling using LSTM architecture demonstrate that the AdamSSM algorithm improves the gap between generalization accuracy and faster convergence than the recent adaptive gradient methods. [ABSTRACT FROM AUTHOR]

Recent years has witnessed the success of convolutional neural networks (CNNs) in many machine learning and pattern recognition applications, especially in image recognition. However, due to the increasing model complexity, the parameter redundancy problem arises, and greatly degrades the performance of CNNs. To alleviate this problem, various regularization techniques, such as Dropout, have been proposed and proved their effectiveness. In this paper, we propose a novel adaptive kernel-based weight decorrelation (AKWD) framework, in order to regularize CNNs for better generalization. Different from existing works, the correlation between paring weights is measured by the cosine distance defined in RKHS associated with a specific kernel. The case with the well-known Gaussian kernel is investigated in detail, where the bandwidth parameter is adaptively estimated. By regularizing CNN models of different capacities using AKWD, better performance is achieved on several benchmark databases for both object classification and face verification tasks. In particular, when Dropout or BatchNorm is present, even higher improvements are obtained using the proposed AKWD, that demonstrates a good compatibility of the proposed regularizer with other regularization techniques. [ABSTRACT FROM AUTHOR]

Deep neural networks (DNNs) display good performance in the domains of recognition and prediction, such as on tasks of image recognition, speech recognition, video recognition, and pattern analysis. However, adversarial examples, created by inserting a small amount of noise into the original samples, can be a serious threat because they can cause misclassification by the DNN. Adversarial examples have been studied primarily in the context of images, but their effect in the audio context is now drawing considerable interest as well. For example, by adding a small distortion to an original audio sample, imperceptible to humans, an audio adversarial example can be created that humans hear as error-free but that causes misunderstanding by a machine. Therefore, it is necessary to create a method of defense for resisting audio adversarial examples. In this paper, we propose an acoustic-decoy method for detecting audio adversarial examples. Its key feature is that it adds well-formalized distortions using audio modification that are sufficient to change the classification result of an adversarial example but do not affect the classification result of an original sample. Experimental results show that the proposed scheme can detect adversarial examples by reducing the similarity rate for an adversarial example to 6.21%, 1.27%, and 0.66% using low-pass filtering (with 12 dB roll-off), 8-bit reduction, and audio silence removal techniques, respectively. It can detect an audio adversarial example with a success rate of 97% by performing a comparison with the initial audio sample. [ABSTRACT FROM AUTHOR]

Deep learning is a branch of neural network which has been intensively developed in the last decade. Due to the high-accuracy classification ability, the deep learning algorithms have been widely used in many fields, such as speech recognition, image recognition, and natural speech processing. However, they also show some shortcomings especially on the selection of some parameters in the network, including hyper-parameters, which is still treated as a time consuming task. In this paper, a modified ABC (ABC-ISB) optimization algorithm is proposed to automatically train the parameters of Feed-Forward Artificial Neural Networks, which is a typical a neural network. In the proposed ABC algorithm, we utilize the information of neighbors with better performance to accelerate the convergence of employed and onlooker bees respectively. In addition, a new selection strategy and a gbest-guided strategy are introduced to enhance the global search capability and balance the exploration and exploitation of the algorithm separately. The experimental results show our ABC-ISB is generally leading and competitive. [ABSTRACT FROM AUTHOR]