Major Depressive Disorder (MDD) is a serious and highly heterogeneous psychological disorder. According to the network hypothesis, depression originates from abnormal neural network information processing, typically resulting in aberrant changes ...

• A novel multilayer network fused with multi-domain information of high-density EEG sensors is proposed.
• The frequency- and temporal-domain topological properties describing information segregation and integration are developed.
• ...

Image colorization estimates RGB colors for grayscale images or video frames to improve their aesthetic and perceptual quality. Over the last decade, deep learning techniques for image colorization have significantly progressed, necessitating a ...

• We review colorization, providing settings, metrics, datasets and comparison.
• We propose a novel colorization network’s taxonomy on type, structure, I/O etc.
• We present a new colorization benchmark dataset, the Natural-Color ...

Multi-view learning (MVL) is a promising data fusion technique based on the principles of consensus and complementarity. Despite significant advancements in this field, several challenges persist. First, scalability remains an issue, as many ...

• MVIKT is specifically designed for classifying multiple views.
• Propose a novel interactive strategy to implement consensus and complementarity.
• Provide a theoretical analysis for the effectiveness of the proposed strategy.
• ...

Biometric authentication has garnered significant attention as a secure and efficient method of identity verification. Among the various modalities, hand vein biometrics, including finger vein, palm vein, and dorsal hand vein recognition, offer ...

• Hand vein biometrics provide high accuracy, low forgery risk, and non-intrusiveness.
• Review explores DL techniques for FV, PV, DHV, and multimodal vein recognition.
• Covers datasets, SOTA metrics, and best-performing vein ...

Amidst the rapid propagation of multimodal fake news across various social media platforms, the identification and filtering of disinformation have emerged as critical areas of academic research. A salient characteristic of fake news lies in its ...

• We explore knowledge-aware multimodal pre-training for fake news detection.
• We propose KAMP, a model for beneficial signals from multiple modalities.
• Our model consistently outperforms SOTA fake news detection baselines in tests.

Recently, convolutional neural networks (CNNs) and multiple instance learning (MIL) methods have been successfully applied to MRI images. However, CNNs directly utilize the whole image as the model input and the downsampling strategy (like max or ...

• We introduce a dual-branch network with cross-attention for liver tumor-related classification.
• We design a cross-attention module between two branches for interpreting lesion-relevant regions.
• We propose a novel loss function to ...

Pragmatics, situated in the domains of linguistics and computational linguistics, explores the influence of context on language interpretation, extending beyond the literal meaning of expressions. It constitutes a fundamental element for natural ...

• A comprehensive survey of five significant pragmatic processing tasks.
• A structured review covering theoretical research, methodology, and applications.
• A systematic summary of technical and application trends, and future ...

Multimodal Emotion Recognition (MER) involves integrating information of various modalities, including audio, visual, text and physiological signals, to comprehensively grasp human sentiments, which has emerged as a vibrant area within human–...

• A novel unsupervised Multiplex Graph Aggregation and Feature Refinement framework.
• Cross-modal Multiplex Graph Structure Aggregation for multimodal fusion.
• Dynamic Feature Refinement and Cross-modal Alignment and Enhancement for ...

A Bayes filter is a widely used estimation algorithm, but it has inherent limitations. Performance can degrade when the dynamics are highly nonlinear or when the probability distribution of the state is unknown. To mitigate these issues, machine ...

• This paper reviews approaches that integrate Bayes filters with machine learning.
• This provides readers with a comprehensive understanding of the current state of research.
• The review systematically categorises the approaches and ...

Complete traffic data collection is crucial for intelligent transportation system, but due to various factors such as cost, it is not possible to deploy sensors at every location. Using spatial interpolation, the traffic data for unobserved ...

• Designed a Dynamic Graph Learning (DGL) module based on self-attention mechanism.
• Proposed a new dual branch architecture to model the diffusion mechanism among nodes.
• Explored a novel auxiliary branch to model the local details of ...

Complementary Label Learning (CLL), a crucial aspect of weakly supervised learning, has seen significant theoretical and practical advancements. However, a comprehensive review of the field has been lacking. This survey provides the first ...

Display Omitted

• The first review of complementary label learning.
• A novel categorization of CLL algorithms.
• Comprehensive experimental evaluation of SOTA methods.
• Delve into various extensions of complementary label learning.
• Elucidate the ...

Recently, the text-to-image diffusion model has gained considerable attention from the community due to its exceptional image generation capability. A representative model, Stable Diffusion, amassed more than 10 million users within just two ...

• We highlight that the text-to-image diffusion model has vulnerabilities in both robustness and safety.
• We provide an in-depth analysis of adversarial attacks on the text-to-image diffusion model.
• We present a detailed analysis of ...

Classifier fusion is established as an effective methodology for boosting performance in different classification settings and one-class classification is no exception. In this study, we consider the one-class classifier fusion problem by ...

• Addressing the one-class classifier fusion by modelling ensemble sparsity.
• Formulation of the learning task as a constrained convex optimisation task.
• Effectively solving the constrained convex optimisation problem.

Nowadays, deep learning depends on large-scale labeled datasets, which limits its broader application in electrocardiogram (ECG) analysis, as manual labeling of ECGs is consistently costly. To overcome this issue, this paper proposes a fused self-...

• An LFBT fuses information from multi-lead ECGs in Self-Supervised Learning.
• A Fused BT loss helps generate high-quality representations for multi-lead ECGs.
• LFBT makes the model achieve better results when training data are scarce.

Transformer architectures have become increasingly popular in healthcare applications. Through outstanding performance in natural language processing and superior capability to encode sequences, transformers have influenced researchers from ...

• Transformers and the attention mechanism for time-series analysis.
• Physiological events detection from biosignals using transformers.
• Comprehensive summary of transformer studies across various biosignal categories.
• Challenges ...

Recent advancements in multi-view unsupervised feature selection (MUFS) have been notable, yet two primary challenges persist. First, real-world datasets frequently consist of unbalanced incomplete multi-view data, a scenario not adequately ...

• The first unbalanced incomplete multi-view feature selection model is designed.
• Tensor low-rank representation learning achieves missing sample recovery.
• Considering errors in self-representation tensor enhances model’s robustness.

In Natural Language Processing (NLP), state-of-the-art machine learning models heavily depend on vast amounts of training data. Often, this data is sourced from third parties, such as crowdsourcing platforms, to enable swift and efficient ...

• NLP models are vulnerable to malicious poisoning attacks.
• Current defenses against attacks are limited and often costly.
• Personalized NLP architectures bolster defense against these threats.
• User-ID excels in protecting ...

Next-item recommendation aims to predict users’ subsequent behaviors using their historical sequence data. However, sessions are often anonymous, short, and time-varying, making it challenging to capture accurate and evolving item ...

• An evolving graph fusion framework is proposed to improve next item recommendation.
• Intra-session graph is utilized for capturing user’s in-sequence interest.
• Evolving session embeddings are estimated with attention mechanism.
• ...

Machine learning methods in healthcare have traditionally focused on using data from a single modality, limiting their ability to effectively replicate the clinical practice of integrating multiple sources of information for improved decision ...

• Review of multimodal machine learning approaches in healthcare.
• Evaluation of common data fusion techniques.
• Overview of multimodal deep-learning framework for model development.

• A temporal-spatial-spectral network for high-resolution land cover segmentation.
• A new benchmark with dense time-series and high-spatial hyperspectral images.
• A fusion module to reduce spatial-resolution differences between multi-...

Recently, deep learning models have found extensive application in high-resolution land-cover segmentation research. However, the most current research still suffers from issues such as insufficient utilization of multi-modal information, which ...