Search

Incomplete multi-view clustering (IMVC) aims to partition samples into different groups for datasets with missing samples. The primary goal of IMVC is to effectively address the challenge posed by missing information in clustering analysis. Most existing IMVC methods focus on balanced incomplete multi-view data, assuming a uniform missing rate across all views. However, this assumption does not accurately reflect real-life scenarios. In reality, unbalanced incomplete multi-view data, characterized by varying missing rates among different views, is more prevalent. This presents significant challenges to the clustering process, as varying missing rates can lead to information imbalance. To address these challenges, this paper introduces a novel approach called projected cross-view learning for unbalanced incomplete multi-view clustering (PCL_UIMVC). Specifically, a reconstruction term is integrated, which leverages the information from the existing samples to facilitate the completion of the unbalanced incomplete multi-view data. Next, a projection matrix is incorporated into the model to harmonize feature dimensions across views, mitigating the impact of information imbalance. Then, a graph regularization term is integrated to preserve the geometric structure of the original data. Finally, an iterative algorithm is developed to solve the proposed model. Extensive experiments on eight standard datasets, featuring various rates of missing data, validate the superior clustering performance of the proposed method.

Data augmentation has contributed to the rapid advancement of unsupervised learning on 3D point clouds. However, we argue that data augmentation is not ideal, as it requires a careful application-dependent selection of the types of augmentations to be performed, thus potentially biasing the information learned by the network during self-training. Moreover, several unsupervised methods only focus on uni-modal information, thus potentially introducing challenges in the case of sparse and textureless point clouds. To address these issues, we propose an augmentation-free unsupervised approach for point clouds, named CluRender, to learn transferable point-level features by leveraging uni-modal information for soft clustering and cross-modal information for neural rendering. Soft clustering enables self-training through a pseudo-label prediction task, where the affiliation of points to their clusters is used as a proxy under the constraint that these pseudo-labels divide the point cloud into approximate equal partitions. This allows us to formulate a clustering loss to minimize the standard cross-entropy between pseudo and predicted labels. Neural rendering generates photorealistic renderings from various viewpoints to transfer photometric cues from 2D images to the features. The consistency between rendered and real images is then measured to form a fitting loss, combined with the cross-entropy loss to self-train networks. Experiments on downstream applications, including 3D object detection, semantic segmentation, classification, part segmentation, and few-shot learning, demonstrate the effectiveness of our framework in outperforming state-of-the-art techniques.

This paper aims to craft adversarial queries for image retrieval, which uses image features for similarity measurement. Many commonly used methods are developed in the context of image classification. However, these methods, which attack prediction probabilities, only exert an indirect influence on the image features and are thus found less effective when being applied to the retrieval problem. In designing an attack method specifically for image retrieval, we introduce opposite-direction feature attack (ODFA), a white-box attack approach that directly attacks query image features to generate adversarial queries. As the name implies, the main idea underpinning ODFA is to impel the original image feature to the opposite direction, similar to a U-turn. This simple idea is experimentally evaluated on five retrieval datasets. We show that the adversarial queries generated by ODFA cause true matches no longer to be seen at the top ranks, and the attack success rate is consistently higher than classifier attack methods. In addition, our method of creating adversarial queries can be extended for multi-scale query inputs and is generalizable to other retrieval models without foreknowing their weights, i.e., the black-box setting.

3D human reconstruction is an important technology connecting the real world and the virtual world, but most of previous work needs expensive computing resources, making it difficult in real-time scenarios. We propose a lightweight human body reconstruction system based on parametric model, which employs only one RGBD camera as input. To generate a human model end to end, we build a fast and lightweight deep-learning network named Fast Body Net (FBN). The network pays more attention on the face and hands to enrich the local details. Additionally, we train a denoising auto-encoder to reduce unreasonable states of human model. Due to the lack of human dataset based on RGBD images, we propose an Indoor-Human dataset to train the network, which contains a total of 2500 frames of action data of five actors collected by Azure Kinect camera. Depth images avoid using RGB to extract depth features, which makes FBN lightweight and high-speed in reconstructing parametric human model. Qualitative and quantitative analysis on experimental results show that our method can improve at least 57% in efficiency with similar accuracy, as compared to state-of-the-art methods. Through our study, it is also demonstrated that consumer-grade RGBD cameras can provide great applications in real-time display and interaction for virtual reality.

This paper studies compositional 3D-aware image synthesis for both single-object and multi-object scenes. We observe that two challenges remain in this field: existing approaches (1) lack geometry constraints and thus compromise the multi-view consistency of the single object, and (2) can not scale to multi-object scenes with complex backgrounds. To address these challenges coherently, we propose multi-view consistent generative adversarial networks (MVCGAN) for compositional 3D-aware image synthesis. First, we build the geometry constraints on the single object by leveraging the underlying 3D information. Specifically, we enforce the photometric consistency between pairs of views, encouraging the model to learn the inherent 3D shape. Second, we adapt MVCGAN to multi-object scenarios. In particular, we formulate the multi-object scene generation as a “decompose and compose” process. During training, we adopt the top-down strategy to decompose training images into objects and backgrounds. When rendering, we deploy a reverse bottom-up manner by composing the generated objects and background into the holistic scene. Extensive experiments on both single-object and multi-object datasets show that the proposed method achieves competitive performance for 3D-aware image synthesis.

Automated Breast Ultrasound is a highly advanced breast tumor detection modality that produces hundreds of 2D slices in each scan. However, this large number of slices poses a significant burden for physicians to review. This paper proposes a novel 2.5D tumor detection model, named “ABUSDet,” to assist physicians in automatically reviewing ABUS images and predicting the locations of breast tumors in images. At the core of this approach, a sequence of data blocks partitioned from a pre-processed 3D volume are fed to a 2.5D tumor detection model, which outputs a sequence of 2D tumor candidates. An aggregation module then clusters the 2D tumor candidates to produce the ultimate 3D coordinates of the tumors. To further improve the accuracy of the model, a novel mechanism for training deep learning models, called “Deliberate Training,” is proposed. The proposed model is trained and tested on a dataset of 87 patients with 235 ABUS volumes. It achieves sensitivities of 77.94%, 75.49%, and 65.19% at FPs/volume of 3, 2, and 1, respectively. Compared with the 2D and 3D object detection models, the proposed ABUSDet model achieves the highest sensitivity with relatively low false-positive rates. Graphical abstract: [Figure not available: see fulltext.]

In the paradigm of online continual learning, one neural network is exposed to a sequence of tasks, where the data arrive in an online fashion and previously seen data are not accessible. Such online fashion causes insufficient learning and severe forgetting on past tasks issues, preventing a good stability-plasticity trade-off, where ideally the network is expected to have high plasticity to adapt to new tasks well and have the stability to prevent forgetting on old tasks simultaneously. To solve these issues, we propose a trust-region adaptive frequency approach, which alternates between standard-process and intra-process updates. Specifically, the standard-process replays data stored in a coreset and interleaves the data with current data, and the intra-process updates the network parameters based on the coreset. Furthermore, to improve the unsatisfactory performance stemming from online fashion, the frequency of the intra-process is adjusted based on a trust region, which is measured by the confidence score of current data. During the intra-process, we distill the dark knowledge to retain useful learned knowledge. Moreover, to store more representative data in the coreset, a confidence-based coreset selection is presented in an online manner. The experimental results on standard benchmarks show that the proposed method significantly outperforms state-of-art continual learning algorithms.

Data science has been the foundation of recommender systems for a long time. Over the past few decades, various recommender systems have been developed using different data science and machine learning methodologies and techniques. However, no existing work systematically discusses the significant relationships between data science and recommender systems. To bridge this gap, this paper aims to systematically investigate recommender systems from the perspective of data science. Firstly, we introduce the various types of data used for recommendations and the corresponding machine learning models and methods that effectively represent each type. Next, we provide a brief outline of the representative data science and machine learning models utilized in building recommender systems. Subsequently, we share some preliminary thoughts on next-generation recommender systems. Finally, we summarize this special issue on data science for next-generation recommender systems.

Visual-language navigation (VLN) is the task of entailing an agent to carry out navigational instructions inside photo-realistic environments. One of the key challenges in VLN is how to conduct robust navigation by mitigating the uncertainty caused by ambiguous instructions and insufficient observation of the environment. Agents trained by current approaches typically suffer from this and would consequently struggle to take navigation actions at every step. In contrast, when humans face such a challenge, we can still maintain robust navigation by actively exploring the surroundings to gather more information and thus make a more confident navigation decision. This work draws inspiration from human navigation behavior and endows an agent with an active perception ability for more intelligent navigation. To achieve this, we propose an end-to-end framework for learning an exploration policy that decides (i) when and where to explore, (ii) what information is worth gathering during exploration, and (iii) how to adjust the navigation decision after the exploration. In this way, the agent is able to turn its past experiences as well as new explored knowledge to contexts for more confident navigation decision making. In addition, an external memory is used to explicitly store the visited visual environments and thus allows the agent to adopt a late action-taking strategy to avoid duplicate exploration and navigation movements. Our experimental results on two standard benchmark datasets show promising exploration strategies emerged from training, which leads to significant boost in navigation performance.

Correlation coefficient (CC) is a reliable information measure for measuring interrelationship between Pythagorean fuzzy sets (PFSs). Some approaches for calculating CC of PFSs have been considered. These hitherto approaches assess only the strength of relationship between PFSs, and are described within the interval [0,1]. This paper proposes a three-way approach for the computation of CC between PFSs by using the concepts of variance and covariance, respectively. This new approach is defined within the interval [− 1,1] akin to classical statistics, shows the strength of relationship between the considered PFSs and indicates whether the PFSs are either positively or negatively correlated. By including the three conventional parameters of PFSs in the proposed technique, the possibility of error due to information leakage is reasonably minimized. The new technique is validated with some theoretical results to show its suitability as reliable information measure. Some numerical examples are considered to show the edges of the new methods over similar methods. From the comparative analysis, the proposed methods of computing CCPFSs give more reliable and reasonable results compare to similar existing methods as presented in Table 13. Certain decision-making problems involving recognition of patterns and diagnostic medicine are resolved with the aid of the new method. The three-way technique of computing correlation coefficient between PFSs can solve decision-making problems that are multi-attributes in nature.

The uncertain world has seen increasing emergencies, crises and disasters (ECDs), such as the COVID-19 pandemic, hurricane Ian, global financial inflation and recession, misinformation disaster, and cyberattacks. AI for smart disaster resilience (AISDR) transforms classic reactive and scripted disaster management to digital proactive and intelligent resilience across ECD ecosystems. A systematic overview of diverse ECDs, classic ECD management, ECD data complexities, and an AISDR research landscape are presented in this article. Translational disaster AI is essential to enable smart disaster resilience.

Underwater robots in shallow waters usually suffer from strong wave forces, which may frequently exceed robot’s control constraints. Learning-based controllers are suitable for disturbance rejection control, but the excessive disturbances heavily affect the state transition in Markov Decision Process (MDP) or Partially Observable Markov Decision Process (POMDP). This issue is amplified by training-test model mismatch. In this paper, we propose a transfer reinforcement learning algorithm using Transition Mismatch Compensation (TMC), that learns an additional compensatory policy through minimizing mismatch of transitions predicted by the two dynamics models of the source and target tasks. A modular network of learning policies is applied, composed of a Generalized Control Policy (GCP) and an Online Disturbance Identification Model (ODI). GCP is first trained over a wide array of disturbance waveforms. ODI then learns to use past states and actions of the system to predict the disturbance waveforms which are provided as input to GCP (along with the system state). We demonstrated on a pose regulation task in simulation that TMC is able to successfully reject the disturbances and stabilize the robot under an empirical model of the robot system, meanwhile improve sample efficiency.

Visual Question Answering (VQA) is a multimodal task that requires models to understand both textual and visual information. Various VQA models have applied the Transformer structure due to its excellent ability to model self-attention global dependencies. However, balancing global and local dependency modeling in traditional Transformer structures is an ongoing issue. A Transformer-based VQA model that only models global dependencies cannot effectively capture image context information. Thus, this paper proposes a novel Local Self-Attention in Transformer (LSAT) for a visual question answering model to address these issues. The LSAT model simultaneously models intra-window and inter-window attention by setting local windows for visual features. Therefore, the LSAT model can effectively avoid redundant information in global self-attention while capturing rich contextual information. This paper uses grid visual features to conduct extensive experiments and ablation studies on the VQA benchmark datasets VQA 2.0 and CLEVR. The experimental results show that the LSAT model outperforms the benchmark model in all indicators when the appropriate local window size is selected. Specifically, the best test results of LSAT using grid visual features on the VQA 2.0 and CLEVR datasets were 71.94% and 98.72%, respectively. Experimental results and ablation studies demonstrate that the proposed method has good performance. Source code is available at https://github.com/shenxiang-vqa/LSAT.

We introduce HexBox, an intuitive modeling method and interactive tool for creating and editing hexahedral meshes. Hexbox brings the major and widely validated surface modeling paradigm of surface box modeling into the world of hex meshing. The main idea is to allow the user to box-model a volumetric mesh by primarily modifying its surface through a set of topological and geometric operations. We support, in particular, local and global subdivision, various instantiations of extrusion, removal, and cloning of elements, the creation of non-conformal or conformal grids, as well as shape modifications through vertex positioning, including manual editing, automatic smoothing, or, eventually, projection on an externally-provided target surface. At the core of the efficient implementation of the method is the coherent maintenance, at all steps, of two parallel data structures: a hexahedral mesh representing the topology and geometry of the currently modeled shape, and a directed acyclic graph that connects operation nodes to the affected mesh hexahedra. Operations are realized by exploiting recent advancements in grid-based meshing, such as mixing of 3-refinement, 2-refinement, and face-refinement, and using templated topological bridges to enforce on-the-fly mesh conformity across pairs of adjacent elements. A direct manipulation user interface lets users control all operations. The effectiveness of our tool, released as open source to the community, is demonstrated by modeling several complex shapes hard to realize with competing tools and techniques.

Sensor techniques and emerging CNN models have greatly facilitated the development of collaborative fault diagnosis. Existing CNN models apply different fusion schemes to achieve reliable fault identification based on multisensor data. Few CNN models, however, take into account both the intrinsic correlations and the distribution gap between different signals, which may result in a limited exploration of multisource data. To address this issue, a novel convolutional fusion framework called a collaborative fusion convolutional neural network (CFCNN) is developed in this paper. More specifically, a multiscale shrinkage denoising module (MSDM) is developed first to extract multilevel modality-specific features from different mechanical signals. Then, drawing inspiration from the intermediate fusion scheme, a central fusion module (CFM) is introduced to explore the intrinsic correlations and integrate cross-modal features. Moreover, an online label smoothing training (OLST) strategy is applied to reduce overfitting and promote better classification performance of CFCNN. The developed CFCNN is expected to shed new light on collaborative fault diagnosis using the intermediate fusion scheme. The efficacy of the developed CFCNN is verified through the cylindrical rolling bearing dataset and the planetary gearbox dataset.

As online social media platforms continue to proliferate, users are faced with an overwhelming amount of information, making it challenging to filter and locate relevant information. While personalized recommendation algorithms have been developed to help, most existing models primarily rely on user behavior observations such as viewing history, often overlooking the intricate connection between the reading content and the user’s prior knowledge and interest. This disconnect can consequently lead to a paucity of diverse and personalized recommendations. In this paper, we propose a novel approach to tackle the multifaceted issue of recommendation. We introduce the Dual-Observation-based approach for the Recommendation (DOR) system, a novel model leveraging dual observation mechanisms integrated into a deep neural network. Our approach is designed to identify both the core theme of an article and the user’s unique engagement with the article, considering the user’s belief network, i.e., a reflection of their personal interests and biases. Extensive experiments have been conducted using real-world datasets, in which the DOR model was compared against a number of state-of-the-art baselines. The experimental results explicitly demonstrate the reliability and effectiveness of the DOR model, highlighting its superior performance in news recommendation tasks.

Recent days witness significant progress in various multi-modal tasks made by Contrastive Language-Image Pre-training (CLIP), a multi-modal large-scale model that learns visual representations from natural language supervision. However, the potential effects of CLIP on cross-modal hashing retrieval has not been investigated yet. In this paper, we for the first time explore the effects of CLIP on cross-modal hashing retrieval performance and propose a simple but strong baseline Unsupervised Contrastive Multi-modal Fusion Hashing network (UCMFH). We first extract the off-the-shelf visual and linguistic features from the CLIP model, as the input sources for cross-modal hashing functions. To further mitigate the semantic gap between the image and text features, we design an effective contrastive multi-modal learning module that leverages a multi-modal fusion transformer encoder supervising by a contrastive loss, to enhance modality interaction while improving the semantic representation of each modality. Furthermore, we design a contrastive hash learning module to produce high-quality modal-correlated hash codes. Experiments show that significant performance improvement can be made by our simple new unsupervised baseline UCMFH compared with state-of-the-art supervised and unsupervised cross-modal hashing methods. Also, our experiments demonstrate the remarkable performance of CLIP features on cross-modal hashing retrieval task compared to deep visual and linguistic features used in existing state-of-the-art methods. The source codes for our approach is publicly available at: https://github.com/XinyuXia97/UCMFH.

After the many ups and downs over the past 70 years of AI and 50 years of data science (DS), AI/DS have migrated into their new age. This new-generation AI/DS build on the consilience and universology of science, technology and engineering. In particular, it synergizes AI and data science, inspiring Trans-AI/DS (i.e., Trans-AI, Trans-DS and their hybridization) thinking, vision, paradigms, approaches and practices. Trans-AI/DS feature their transformative (or transformational), transdisciplinary, and translational AI/DS in terms of thinking, paradigms, methodologies, technologies, engineering, and practices. Here, we discuss these important paradigm shifts and directions. Trans-AI/DS encourage big and outside-the-box thinking beyond the classic AI, data-driven, model-based, statistical, shallow and deep learning hypotheses, methodologies and developments. They pursue foundational and original AI/DS thinking, theories and practices from the essence of intelligences and complexities inherent in humans, nature, society, and their creations.

The accurate segmentation of medical images plays a crucial role in diagnosing and treating diseases. Although many methods now use multimodal joint segmentation, the joint use of segmentation features extracted by multiple models can lead to heterogeneity and uncertainty. Unreasonable fusion methods cannot exploit the advantages of multiple models and still lack good performance in segmentation. Therefore, this study proposes the FTransCNN model, which is composed of a convolutional neural network (CNN) and Transformer and is based on a fuzzy fusion strategy that jointly utilizes the features extracted by a CNN and Transformer through a new fuzzy fusion module. First, the CNN and Transformer act as the backbone network for parallel feature extraction. Second, channel attention is used to promote the global key information of Transformer to improve the feature representation ability, and spatial attention is used to enhance the local details of CNN features and suppress irrelevant regions. Third, the proposed model applies the Hadamard product to model fine-grained interactions between the two branches and uses the Choquet fuzzy integral to suppress heterogeneity and uncertainty in fused features. Fourth, FTransCNN employs fuzzy attention fusion module (FAFM) hierarchical upsampling to effectively capture both low-level spatial features and high-level semantic context. Finally, the new model obtains the final segmentation result by using the deconvolution and results in an improvement in segmentation. The experimental results on Chest X-ray and Kvasir-SEG dataset show that FTransCNN has better performance on segmentation tasks than the-state-of-the-art deep segmentation models.

Myocardial infarction (MI) is detected using electrocardiography (ECG) signals. Machine learning (ML) models have been used for automated MI detection on ECG signals. Deep learning models generally yield high classification performance but are computationally intensive. We have developed a novel multilevel hybrid feature extraction-based classification model with low time complexity for MI classification. The study dataset comprising 12-lead ECGs belonging to one healthy and 10 MI classes were downloaded from a public ECG signal databank. The model architecture comprised multilevel hybrid feature extraction, iterative feature selection, classification, and iterative majority voting (IMV). In the hybrid handcrafted feature (HHF) generation phase, both textural and statistical feature extraction functions were used to extract features from ECG beats but only at a low level. A new pooling-based multilevel decomposition model was presented to enable them to create features at a high level. This model used average and maximum pooling to create decomposed signals. Using these pooling functions, an unbalanced tree was obtained. Therefore, this model was named multilevel unbalanced pooling tree transformation (MUPTT). On the feature extraction side, two extractors (functions) were used to generate both statistical and textural features. To generate statistical features, 20 commonly used moments were used. A new, improved symmetric binary pattern function was proposed to generate textural features. Both feature extractors were applied to the original MI signal and the decomposed signals generated by the MUPTT. The most valuable features from among the extracted feature vectors were selected using iterative neighborhood component analysis (INCA). In the classification phase, a one-dimensional nearest neighbor classifier with ten-fold cross-validation was used to obtain lead-wise results. The computed lead-wise results derived from all 12 leads of the same beat were input to t

Background: Epilepsy is one of the most commonly seen neurologic disorders worldwide and has generally caused seizures. Electroencephalography (EEG) is widely used in seizure diagnosis. To detect epilepsy automatically, various machine learning (ML) models have been introduced in the literature, but the used EEG signal datasets for epilepsy detection are relatively small. Our main objective is to present a large EEG signal dataset and investigate the detection ability of a new hypercube pattern-based framework using the EEG signals. Material and method: This study collected a large EEG signal dataset (10,356 EEG signals) from 121 participants. We proposed a new information fusion-based feature engineering framework to get high classification performance from this dataset. The dataset consists of 35 channels, and our proposed feature engineering model extracts features from each channel. A new hypercube-based feature extractor has been proposed to generate two feature vectors in the feature extraction phase. Various statistical parameters of the signals have been used to create a feature vector. Multilevel discrete wavelet transform (MDWT) has been applied to develop a multileveled feature extraction function, and seven feature vectors have been extracted. In this work, we have extracted 245 (=35 × 7) feature vectors, and the most valuable features from these vectors have been selected using the neighborhood component analysis (NCA) selector. Finally, these selected features were fed to the k nearest neighbors (kNN) classifier with the leave one subject out (LOSO) cross-validation (CV) strategy. These results have been voted/fused to obtain the highest classification performance. Results: In this work, we have attained 87.78% classification accuracy using voting these vectors and 79.07% with LOSO CV with the EEG signals. Conclusions: The proposed fusion-based feature engineering model achieved satisfactory classification performance using the largest EEG signal datas

Brain diseases, including tumors and mental and neurological disorders, seriously threaten the health and well-being of millions of people worldwide. Structural and functional neuroimaging modalities are commonly used by physicians to aid the diagnosis of brain diseases. In clinical settings, specialist doctors typically fuse the magnetic resonance imaging (MRI) data with other neuroimaging modalities for brain disease detection. As these two approaches offer complementary information, fusing these neuroimaging modalities helps physicians accurately diagnose brain diseases. Typically, fusion is performed between a functional and a structural neuroimaging modality. Because the functional modality can complement the structural modality information, thus improving the performance for the diagnosis of brain diseases by specialists. However, analyzing the fusion of neuroimaging modalities is difficult for specialist doctors. Deep Learning (DL) is a branch of artificial intelligence that has shown superior performances compared to more conventional methods in tasks such as brain disease detection from neuroimaging modalities. This work presents a comprehensive review paper in the field of brain disease detection from the fusion of neuroimaging modalities using DL models like convolutional neural networks (CNNs), recurrent neural networks (RNNs), pretrained, generative adversarial networks (GANs), and Autoencoders (AEs). First, neuroimaging modalities and the need for fusion are discussed. Then, review papers published in the field of neuroimaging multimodalities using AI techniques are explored. Moreover, fusion levels based on DL methods, including input, layer, and decision, with related studies conducted on diagnosing brain diseases, are discussed. Other sections present the most important challenges for diagnosing brain diseases from the fusion of neuroimaging modalities. In the discussion section, the details of previous research on the fusion of neuroimaging modalit

In recent years, research on algorithmic trading based on machine learning has been increasing. One challenge faced is getting an accurate representation of the stock market environment from multi-type data. Most existing algorithmic trading studies analyze the stock market based on a relatively single data source. However, with the complicated stock market environment, different types of data reflect the changes in the stock market from different perspectives, and how to obtain the temporal features of different types of data and integrate them to obtain a deeper representation of the stock market environment are still problems to be solved. To tackle these problems, in this study, we combine deep learning and reinforcement learning (RL) and propose a multi-type data fusion framework with deep reinforcement learning (MSF-DRL) that integrates stock data, technical indicators and candlestick charts, in which technical indicators can reduce the impact of noise in stock data. In the process of learning trading strategies under the MSF-DRL framework, the temporal features of stock data and technical indicators are extracted through a long short-term memory (LSTM) network, and a convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) are successively used to extract the features of the candlestick chart. The fused features are used as the input of the RL module, which makes trading decisions on this basis. To verify the effectiveness of the MSF-DRL framework, we conducted comparative experiments on datasets composed of Chinese stocks and some stocks of the S&P 500 stock market index. Compared with the other trading strategies, our trading strategy can obtain more profits and a higher Sharpe ratio.

Over the years, Machine Learning models have been successfully employed on neuroimaging data for accurately predicting brain age. Deviations from the healthy brain aging pattern are associated with the accelerated brain aging and brain abnormalities. Hence, efficient and accurate diagnosis techniques are required to elicit accurate brain age estimations. Several contributions have been reported in the past for this purpose, resorting to different data-driven modeling methods. Recently, deep neural networks (also referred to as deep learning) have become prevalent in manifold neuroimaging studies, including brain age estimation. In this review, we offer a comprehensive analysis of the literature related to the adoption of deep learning for brain age estimation with neuroimaging data. We detail and analyze different deep learning architectures used for this application, pausing at research works published to date quantitatively exploring their application. We also examine different brain age estimation frameworks, comparatively exposing their advantages and weaknesses. Finally, the review concludes with an outlook towards future directions that should be followed by prospective studies. The ultimate goal of this paper is to establish a common and informed reference for newcomers and experienced researchers willing to approach brain age estimation by using deep learning models.

Narrative visualization is characterized by the integration of data visualization and storytelling techniques. These characteristics provide challenges in its evaluation. Little is known about how these evaluation challenges are addressed by narrative visualization practitioners. We surveyed experienced narrative visualization practitioners to investigate their methods of evaluation. To gain deeper insight we conducted a series of semi-structured interviews with practitioners. We found that there is usually an informal approach to narrative visualization evaluation, where practitioners rely on prior experience and their peers for evaluation. Our study also revealed novel approaches to evaluation. We introduce a practice-led heuristic framework to aid practitioners to evaluate narrative visualization systematically. Our practice-led heuristic framework couples first-hand practitioner experience with recent research literature. This work sheds light on how to address narrative visualization evaluation to better inform both academic research and practice.

Mental health is a basic need for a sustainable and developing society. The prevalence and financial burden of mental illness have increased globally, and especially in response to community and worldwide pandemic events. Children suffering from such mental disorders find it difficult to cope with educational, occupational, personal, and societal developments, and treatments are not accessible to all. Advancements in technology have resulted in much research examining the use of artificial intelligence to detect or identify characteristics of mental illness. Therefore, this paper presents a systematic review of nine developmental and mental disorders (Autism spectrum disorder, Attention deficit hyperactivity disorder, Schizophrenia, Anxiety, Depression, Dyslexia, Post-traumatic stress disorder, Tourette syndrome, and Obsessive–compulsive disorder) prominent in children and adolescents. Our paper focuses on the automated detection of these developmental and mental disorders using physiological signals. This paper also presents a detailed discussion on signal analysis, feature engineering, and decision-making with their advantages, future directions and challenges on the papers published on mental disorders of children. We have presented the details of the dataset description, validation techniques, features extracted and decision-making models. The challenges and future directions present open research questions on signal or availability, uncertainty, explainability, and hardware implementation resources for signal analysis and machine or deep learning models. Finally, the main findings of this study are presented in the conclusion section.

COVID-19, chronic obstructive pulmonary disease (COPD), heart failure (HF), and pneumonia can lead to acute respiratory deterioration. Prompt and accurate diagnosis is crucial for effective clinical management. Chest X-ray (CXR) and chest computed tomography (CT) are commonly used for confirming the diagnosis, but they can be time-consuming and biased. To address this, we developed a computationally efficient deep feature engineering model called Hybrid-Patch-Alex for automated COVID-19, COPD, and HF diagnosis. We utilized one CXR dataset and two CT image datasets, including a newly collected dataset with four classes: COVID-19, COPD, HF, and normal. Our model employed a hybrid patch division method, transfer learning with pre-trained AlexNet, iterative neighborhood component analysis for feature selection, and three standard classifiers (k-nearest neighbor, support vector machine, and artificial neural network) for automated classification. The model achieved high accuracy rates of 99.82%, 92.90%, and 97.02% on the respective datasets, using kNN and SVM classifiers.

Deep learning based stereo matching algorithms have been extensively researched in areas such as robot vision and autonomous driving due to their promising performance. However, these algorithms require a large amount of labeled data for training and encounter inadequate domain adaptability, which degraded their applicability and flexibility. This work addresses the two deficiencies and proposes a few-shot trained stereo matching model with high domain adaptability. In the model, stereo matching is formulated as the problem of dynamic optimization in the possible solution space, and a multi-scale matching cost computation method is proposed to obtain the possible solution space for the application scenes. Moreover, an adaptive recurrent 3D convolutional neural network is designed to determine the optimal solution from the possible solution space. Experimental results demonstrate that the proposed model outperforms the state-of-the-art stereo matching algorithms in terms of training requirements and domain adaptability.