Artificial Intelligence Review最新文献

Graph neural networks (GNNs) process the graph-structured data using neural networks and have proven successful in various graph processing tasks. Currently, graph pooling operators have emerged as crucial components that bridge the gap between node representation learning and diverse graph-level tasks by transforming node representations into graph representations. Given the rapid growth and widespread adoption of graph pooling, this review aims to summarize the existing graph pooling operators for GNNs and their representative applications in omics. Specifically, we first present a comprehensive taxonomy of existing graph pooling algorithms, expanding the categorization for both global and hierarchical pooling operators, and for the first time reviewing the inverse operation of graph pooling, named unpooling. Next, we describe the general evaluation framework for graph pooling operators, encompassing three fundamental aspects: experimental setup, ablation analysis, and model interpretation. We also discuss open issues that significantly influence the design of graph pooling operators, including complexity, connectivity, adaptability, additional loss, and attention mechanisms. Finally, we summarize bioinformatics applications of graph pooling operators in omics, including graphs of gene interaction, medical images, and protein structures for drug discovery and disease diagnosis. Furthermore, we showcase the impact of graph pooling operators on research in specific real-world domains, with a focus on prediction performance and model interpretability. This review provides methodological insights in machine learning based graph modeling and related omics research, as well as an ongoing resource by gathering related papers and code in a dedicated GitHub repository (https://github.com/Hou-WJ/Graph-Pooling-Operators-and-Bioinformatics-Applications).

As a popular meta-heuristic algorithm, the Moth-Flame Optimization (MFO) algorithm has garnered significant interest owing to its high flexibility and straightforward implementation. However, when addressing engineering constraint problems with specific parameters, MFO also exhibits limitations such as fast convergence and a tendency to converge to local optima. In order to address these challenges, this paper introduces an enhanced version of the MFO, EQDXMFO. EQDXMFO integrates a Quality Enhancement (EQ) strategy and a Directional Crossover (DX) mechanism, fortifying the algorithm’s search dynamics. Specifically, the DX mechanism is designed to augment the population’s diversity, enhancing the algorithm’s exploratory potential. Concurrently, the EQ strategy is employed to elevate the solution quality, which in turn refines the convergence precision of the algorithm. To verify the effectiveness of EQDXMFO, experiments are carried out on the test set of the IEEE CEC2017. A total of 5 classical algorithms, five excellent MFO variants, and seven state-of-the-art algorithms are selected for comparison, which confirm the significant advantages of EQDXMFO. Next, EQDXMFO is applied to five complex engineering constraint problems, demonstrating that EQDXMFO can optimize realistic problems. The comprehensive analysis shows that EQDXMFO has strong optimization capabilities and provides methods for research on other complex real-world problems.

Automatic surgical workflow analysis, which aims to recognize the ongoing surgical events in videos, is fundamental for developing context-aware computer-assisted systems. This paper reviews representative surgical workflow recognition algorithms based on deep learning, outlining their merits, limitations, and future research directions. The literature survey was performed on three large bibliographic databases, covering 67 lary sources, which were comparatively analyzed in terms of spatial feature modeling, spatio-temporal feature modeling, input pre-processing, regularization and post-processing algorithms, as well as learning strategies. Then, common public datasets and evaluation metrics for surgical workflow recognition are also described in detail. Finally, we discuss all literature from different perspectives, and point out the challenges, possible solutions and future trends. The need for more diverse and larger datasets, the potential of unsupervised and semi-supervised learning approaches, comprehensive and equitable metrics, establishing complete regulatory and data standards, and interoperability will be key challenges in translating models to clinical operating rooms. And we propose that surgical activity anticipation and employing large language model as training assistant are interesting research directions in surgical workflow analysis.

Over the past few decades, numerous studies have focused on identifying and recognizing human actions using machine learning and computer vision techniques. Video-based human action recognition (HAR) aims to detect actions from video sequences automatically. This can cover simple gestures to complex actions involving multiple people interacting with objects. Actions in team sports exhibit a different nature compared to other sports, since they tend to occur at a faster pace and involve more human-human interactions. As a result, research has typically not focused on the challenges of HAR in team sports. This paper comprehensively summarises HAR-related research and applications with specific focus on team sports such as football (soccer), basketball and Australian rules football. Key datasets used for HAR-related team sports research are explored. Finally, common challenges and future work are discussed, and possible research directions identified.

Network clustering is a critical task in data analysis, aimed at uncovering the underlying structure and patterns within complex networks. Traditional clustering methods often struggle with large-scale and noisy data, leading to suboptimal results. Also, the efficiency of positive samples in network clustering depends on the carefully constructed data augmentation, and the pre-training process of the model deals with large-scale data. To address these issues, in this paper, we introduce an efficient network clustering approach that leverages Graph-Boosting and Nonnegative Matrix Factorization to enhance clustering performance (GBNMF). Our algorithm addresses the limitations of traditional clustering techniques by incorporating the strengths of graph-boosting, which iteratively improves the quality of clusters, and Nonnegative Matrix Factorization (NMF), which effectively captures latent structures within the data. We validate our algorithm through extensive experiments on various benchmark network datasets, demonstrating significant improvements in clustering accuracy and robustness. The proposed algorithm not only achieves superior clustering results but also exhibits remarkable computational efficiency, making it a valuable tool for large-scale network analysis applications.

Generative AI (GenAI) represents a shift from AI’s ability to “recognize” to its ability to “generate” solutions for a wide range of tasks. As generated solutions and applications grow more complex and multi-faceted, new needs, objectives, and possibilities for explainability (XAI) have emerged. This work elaborates on why XAI has gained importance with the rise of GenAI and the challenges it poses for explainability research. We also highlight new and emerging criteria that explanations should meet, such as verifiability, interactivity, security, and cost considerations. To achieve this, we focus on surveying existing literature. Additionally, we provide a taxonomy of relevant dimensions to better characterize existing XAI mechanisms and methods for GenAI. We explore various approaches to ensure XAI, ranging from training data to prompting. Our paper provides a concise technical background of GenAI for non-technical readers, focusing on text and images to help them understand new or adapted XAI techniques for GenAI. However, due to the extensive body of work on GenAI, we chose not to delve into detailed aspects of XAI related to the evaluation and usage of explanations. Consequently, the manuscript appeals to both technical experts and professionals from other fields, such as social scientists and information systems researchers. Our research roadmap outlines over ten directions for future investigation.

Partitional clustering is a type of clustering that can organize the data into non-overlapping groups or clusters. This technique has diverse applications across the different various domains like image processing, pattern recognition, data mining, rule-based systems, customer segmentation, image segmentation, and anomaly detection, etc. Hence, this survey aims to identify the key concepts and approaches in partitional clustering. Further, it also highlights its widespread applicability including major advantages and challenges. Partitional clustering faces challenges like selecting the optimal number of clusters, local optima, sensitivity to initial centroids, etc. Therefore, this survey describes the clustering problems as partitional clustering, dynamic clustering, automatic clustering, and fuzzy clustering. The objective of this survey is to identify the meta-heuristic algorithms for the aforementioned clustering. Further, the meta-heuristic algorithms are also categorised into simple meta-heuristic algorithms, improved meta-heuristic algorithms, and hybrid meta-heuristic algorithms. Hence, this work also focuses on the adoption of new meta-heuristic algorithms, improving existing methods and novel techniques that enhance clustering performance and robustness, making partitional clustering a critical tool for data analysis and machine learning. This survey also highlights the different objective functions and benchmark datasets adopted for measuring the effectiveness of clustering algorithms. Before the literature survey, several research questions are formulated to ensure the effectiveness and efficiency of the survey such as what are the various meta-heuristic techniques available for clustering problems? How to handle automatic data clustering? What are the main reasons for hybridizing clustering algorithms? The survey identifies shortcomings associated with existing algorithms and clustering problems and highlights the active area of research in the clustering field to overcome these limitations and improve performance.

Head pose estimation (HPE) has been extensively investigated over the past decade due to its wide range of applications across several domains of artificial intelligence (AI), resulting in progressive improvements in accuracy. The problem becomes more challenging when the application requires full-range angles, particularly in unconstrained environments, making HPE an active research topic. This paper presents a comprehensive survey of recent AI-based HPE tasks in digital images. We also propose a novel taxonomy based on the main steps to implement each method, broadly dividing these steps into eleven categories under four groups. Moreover, we provide the pros and cons of ten categories of the overall system. Finally, this survey sheds some light on the public datasets, available codes, and future research directions, aiding readers and aspiring researchers in identifying robust methods that exhibit a strong baseline within the subcategory for further exploration in this fascinating area. The review compared and analyzed 113 articles published between 2018 and 2024, distributing 70.5% deep learning, 24.1% machine learning, and 5.4% hybrid approaches. Furthermore, it included 101 articles related to datasets, definitions, and other elements for AI-based HPE systems published over the last two decades. To the best of our knowledge, this is the first paper that aims to survey HPE strategies based on artificial intelligence, with detailed explanations of the main steps to implement each method. A regularly updated project page is provided: (github).

Breast cancer, the earliest documented cancer in history, stands as a foremost cause of mortality, accounting for 684,996 deaths globally in 2020 (15.5% of all female cancer cases). Irrespective of socioeconomic factors, geographic locations, race, or ethnicity, breast cancer ranks as the most frequently diagnosed cancer in women. The standard grading for breast cancer utilizes the Nottingham Histopathology Grading (NHG) system, which considers three crucial features: mitotic counts, nuclear pleomorphism, and tubule formation. Comprehensive reviews on features, for example, mitotic count and nuclear pleomorphism have been available thus far. Nevertheless, a thorough investigation specifically focusing on tubule formation aligned with the NHG system is currently lacking. Motivated by this gap, the present study aims to unravel tubule formation in histopathology images via a comprehensive review of detection approaches involving tubule and tumor features. Without temporal constraints, a structured methodology is established in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, resulting in 12 articles for tubule detection and 67 included articles for tumor detection. Despite the primary focus on breast cancer, the structured search string extends beyond this domain to encompass any cancer type utilizing histopathology images as input, focusing on tubule and tumor detection. This broadened scope is essential. Insights from approaches in tubule and tumor detection for various cancers can be assimilated, integrated, and contributed to an enhanced understanding of tubule formation in breast histopathology images. This study compiles evidence-based analyses into a cohesive document, offering comprehensive information to a diverse audience, including newcomers, experienced researchers, and stakeholders interested in the subject matter.

Deep neural networks (DNNs) have revolutionized artificial intelligence but often lack performance when faced with out-of-distribution data, a common scenario due to the inevitable domain shifts in real-world applications. This limitation stems from the common assumption that training and testing data share the same distribution-an assumption frequently violated in practice. Despite their effectiveness with large amounts of data and computational power, DNNs struggle with distributional shifts and limited labeled data, leading to overfitting and poor generalization across various tasks and domains. Meta-learning presents a promising approach by employing algorithms that acquire transferable knowledge across various tasks for fast adaptation, eliminating the need to learn each task from scratch. This survey paper delves into the realm of meta-learning with a focus on its contribution to domain generalization. We first clarify the concept of meta-learning for domain generalization and introduce a novel taxonomy based on the feature extraction strategy and the classifier learning methodology, offering a granular view of methodologies. Additionally, we present a decision graph to assist readers in navigating the taxonomy based on data availability and domain shifts, enabling them to select and develop a proper model tailored to their specific problem requirements. Through an exhaustive review of existing methods and underlying theories, we map out the fundamentals of the field. Our survey provides practical insights and an informed discussion on promising research directions.