Artificial Intelligence Review最新文献

{"title":"Long-tailed image recognition through balancing discriminant quality","authors":"Yan-Xue Wu,&nbsp;Fan Min,&nbsp;Ben-Wen Zhang,&nbsp;Xian-Jie Wang","doi":"10.1007/s10462-023-10544-x","DOIUrl":"10.1007/s10462-023-10544-x","url":null,"abstract":"<div><p>Long-tailed image recognition is a challenging task in real scenes with large-scale data. Popular strategies, such as loss reweighting and data resampling, aim to reduce the model bias toward head classes. Specifically, different loss reweighting approaches explore various endogenous or exogenous measures. In this paper, we study a new endogenous measure called discriminant quality (DQ) by considering validation accuracy and discriminant uncertainty. DQ takes advantage of continuous information over a period of time. It is more robust than instantaneous information because of the mitigation of measuring instability caused by random perturbations during training. Additionally, the weight of each class is automatically rebalanced based on DQ. Consequently, the class weight supports the design of a dynamic updating strategy for the significance of the DQ difference. Experiments on MNIST-LT, CIFAR-100-LT, ImageNet-LT, and Places-LT demonstrated the superiority of DQ over state-of-the-art ones in terms of prediction accuracy.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"833 - 856"},"PeriodicalIF":12.0,"publicationDate":"2023-07-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45845450","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An improved TOPSIS method for multi-criteria decision making based on hesitant fuzzy β neighborhood","authors":"Chenxia Jin,&nbsp;Jusheng Mi,&nbsp;Fachao Li,&nbsp;Meishe Liang","doi":"10.1007/s10462-023-10510-7","DOIUrl":"10.1007/s10462-023-10510-7","url":null,"abstract":"<div><p>Multi-criteria Decision Making (MCDM) plays a very vital role in many application fields. There are many classical methods to solve the MCDM problems if the available information is crisp. However, the uncertainty and ambiguity inherent in the MCDM often makes these methods unsuitable for solving this kind of problem. Aims at the failures of TOPSIS method that can not rank the alternatives completely in a Hesitant Fuzzy <i>β</i>-Covering Approximation Space (HFβCAS), we develop an improved TOPSIS method. First, we define two pairs of hesitant fuzzy relationship based on hesitant fuzzy <i>β</i>-neighborhood, and construct the corresponding hesitant fuzzy covering rough set models; further we discuss the properties and relationships between the models. Second, we introduce a new comprehensive weight determination method by using the precision degree of hesitant fuzzy covering rough set and the maximizing deviation method. Third, we construct a γ-βCHF-TOPSIS method to MCDM which generalizes the TOPSIS method in an HFβCAS. Finally, two real decision-making problems are used to illustrate the concrete implementation process of γ-βCHF-TOPSIS method, and demonstrate its effectiveness and reasonability.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"793 - 831"},"PeriodicalIF":12.0,"publicationDate":"2023-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"46708054","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-source deep transfer learning algorithm based on feature alignment","authors":"Changhong Ding,&nbsp;Peng Gao,&nbsp;Jingmei Li,&nbsp;Weifei Wu","doi":"10.1007/s10462-023-10545-w","DOIUrl":"10.1007/s10462-023-10545-w","url":null,"abstract":"<div><p>With the deepening of transfer learning research, researchers are no longer satisfied with the classification of knowledge in a single field but hope that the classification of knowledge in multiple fields can be realized, so as to simulate the behavior of human “analogy” and enable the machine to draw inferences”. However, the feature realization of multiple source domains often differs greatly, which brings a challenge to the traditional transfer learning scheme. In this paper, a multi-source deep transfer learning algorithm MDTLFA based on feature alignment is proposed to solve the problem that the data from multiple source domains often has different feature realizations. MDTLFA first reduces the difference in the marginal probability distribution between fields at the sample level by means of the maximum mean deviation MMD. Then, the feature alignment strategy is used at the feature level to further reduce the difference in the marginal probability distribution between the fields and maintain the unique data manifold structure while sharing similar features. On this basis, the conditional probability adaptation CPDA was constructed to reduce the difference in conditional probability distribution between domains and enhance the portability of source domain features. The CPTCNN model was constructed based on a convolutional neural network using CPDA. Finally, the CPTCNN model is trained in the subspace to obtain a classifier set, and the designed strategy is used to select the classifier with a small classification error in the target domain to form MDTLFA. Multiple source domains, marginal probability adaptation at the sample level and feature level, and the CPTCNN model constructed based on the minimization of conditional probability differences effectively improve the performance of data features in multiple domains, thus improving the classification effect. The experimental results on several real data sets show that the MDTLFA algorithm is effective and has some advantages compared with the advanced benchmark algorithm.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"769 - 791"},"PeriodicalIF":12.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41541211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A critical review on applications of artificial intelligence in manufacturing","authors":"Omkar Mypati,&nbsp;Avishek Mukherjee,&nbsp;Debasish Mishra,&nbsp;Surjya Kanta Pal,&nbsp;Partha Pratim Chakrabarti,&nbsp;Arpan Pal","doi":"10.1007/s10462-023-10535-y","DOIUrl":"10.1007/s10462-023-10535-y","url":null,"abstract":"<div><p>The fourth industrial revolution, Industry 4.0, has brought internet, artificial intelligence (AI), and machine learning (ML) concepts into manufacturing. There is an  immediate need to understand the capabilities of AI and ML and how they can be implemented in manufacturing domains. This article presents a detailed survey of AI algorithms and their use in manufacturing. The article treats casting, forming, machining, welding, additive manufacturing (AM), and supply chain management (SCM) as six manufacturing verticals. The horizontals in each vertical are the descriptions including, the evolution of each process from the mechanization era to the present-day scenario, and developments in the automation of processes by processing signal and image information and applying ML and AI algorithms. The evolution of robotics and cloud-based technologies is also discussed. The critical review gives a realistic view of manufacturing automation and benefits of AI. Further, the article discusses several manufacturing use cases where AI and ML algorithms are deployed. As a future research direction, human-like intelligence is introduced highlighting the necessity of cognitive skills in manufacturing. In a nutshell, a reader can logically explain why, when, and how far AI will define complete manufacturing.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"661 - 768"},"PeriodicalIF":12.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41316047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Systematic literature review on identifying influencers in social networks","authors":"Seyed Farid Seyfosadat,&nbsp;Reza Ravanmehr","doi":"10.1007/s10462-023-10515-2","DOIUrl":"10.1007/s10462-023-10515-2","url":null,"abstract":"<div><p>Considering the ever-increasing size and complexity of social networks, developing methods to extract meaningful knowledge and information from users’ vast amounts of data is crucial. Identifying influencers on social networks is one of the essential investigations on these networks and has many applications in marketing, advertising, sociology, behavior analysis, and security issues. In recent years, many studies have been conducted on analyzing and identifying influencers on social networks. Therefore, in this article, a Systematic Literature Review (SLR) has been performed on previous studies about the methods of identifying influencers. To this end, we review the definitions of influencers, the datasets used for evaluation purposes, the methods of identifying influencers, and the evaluation techniques. Furthermore, the quality assessment of the recently published papers also has been performed in different aspects to find whether research about identifying influencers has progressed. Finally, trends and opportunities for future studies about influencers’ identification are presented. The result of this SLR shows that the quantity and quality of articles in the field of identifying influencers in social networks are growing and progressive, which shows this field is a dynamic and active area of research.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"567 - 660"},"PeriodicalIF":12.0,"publicationDate":"2023-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45427221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Battling voice spoofing: a review, comparative analysis, and generalizability evaluation of state-of-the-art voice spoofing counter measures","authors":"Awais Khan,&nbsp;Khalid Mahmood Malik,&nbsp;James Ryan,&nbsp;Mikul Saravanan","doi":"10.1007/s10462-023-10539-8","DOIUrl":"10.1007/s10462-023-10539-8","url":null,"abstract":"<div><p>With the advent of automated speaker verification (ASV) systems comes an equal and opposite development: malicious actors may seek to use voice spoofing attacks to fool those same systems. Various counter measures have been proposed to detect these spoofing attacks, but current offerings in this arena fall short of a unified and generalized approach applicable in real-world scenarios. For this reason, defensive measures for ASV systems produced in the last 6-7 years need to be classified, and qualitative and quantitative comparisons of state-of-the-art (SOTA) counter measures should be performed to assess the effectiveness of these systems against real-world attacks. Hence, in this work, we conduct a review of the literature on spoofing detection using hand-crafted features, deep learning, and end-to-end spoofing countermeasure solutions to detect logical access attacks, such as speech synthesis and voice conversion, and physical access attacks, i.e., replay attacks. Additionally, we review integrated and unified solutions to voice spoofing evaluation and speaker verification, and adversarial and anti-forensic attacks on both voice counter measures and ASV systems. In an extensive experimental analysis, the limitations and challenges of existing spoofing counter measures are presented, the performance of these counter measures on several datasets is reported, and cross-corpus evaluations are performed, something that is nearly absent in the existing literature, in order to assess the generalizability of existing solutions. For the experiments, we employ the ASVspoof2019, ASVspoof2021, and VSDC datasets along with GMM, SVM, CNN, and CNN-GRU classifiers. For reproducibility of the results, the code of the testbed can be found at our GitHub Repository (https://github.com/smileslab/Comparative-Analysis-Voice-Spoofing).</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"513 - 566"},"PeriodicalIF":12.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10462-023-10539-8.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43395522","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AO-HRCNN: archimedes optimization and hybrid region-based convolutional neural network for detection and classification of diabetic retinopathy","authors":"Sujatha Krishnamoorthy,&nbsp;yu Weifeng,&nbsp;Jin Luo,&nbsp;Seifedine Kadry","doi":"10.1007/s10462-023-10516-1","DOIUrl":"10.1007/s10462-023-10516-1","url":null,"abstract":"<div><p>Diabetic Retinopathy (DR) primarily affects a set of lesions in the eyes, causing retinal degeneration and loss of vision. The DR features serve as a crucial component for ophthalmologists to diagnose DR at an earlier stage. This paper presents an automatic DR screening tool using a hybrid GAN-RCNN architecture formulated to categorize and identify different DR grades from the fundus images captured at various resolutions. The hybrid GAN-RCNN architecture is formulated by replacing the discriminator in the GAN with the RCNN classifier. The RCNN model can handle the complex inter-class and intra-class variations present in the fundus retina images and classify them into different classes such as mild, moderate, severe, and nonproliferative DR. The RCNN model not only extracts the pixels present in the fundus image but also focuses on the significant relationship that exists between different DR classes. The Archimedes optimization Algorithm (AOA) is used to optimize the different GAN and RCNN hyperparameters. When compared to the existing techniques the proposed model offers an accuracy of 99%, 98.5%, and 99.4% in the APTOS, Kaggle, and Messidor datasets which is comparatively high. The experimental outcome reveals that the introduced model serves as a concrete baseline for the diagnosis and treatment of DR patients.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"483 - 511"},"PeriodicalIF":12.0,"publicationDate":"2023-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"42459474","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The role of politeness in human–machine interactions: a systematic literature review and future perspectives","authors":"Patrizia Ribino","doi":"10.1007/s10462-023-10540-1","DOIUrl":"10.1007/s10462-023-10540-1","url":null,"abstract":"<div><p>The growing prevalence of interactions between humans and machines, coupled with the rapid development of intelligent and human-like features in technology, necessitates considering the potential implications that an increasingly inter-personal interaction style might have on human behavior. Particularly, since human–human interactions are fundamentally affected by politeness rules, several researchers are investigating if such social norms have some implications also within human–machine interactions. This paper reviews scientific works dealing with politeness issues within human–machine interactions by considering a variety of artificial intelligence systems, such as smart devices, robots, digital assistants, and self-driving cars. This paper aims to analyze scientific results to answer the questions of why technological devices should behave politely toward humans, but above all, why human beings should be polite toward a technological device. As a result of the analysis, this paper wants to outline future research directions for the design of more effective, socially competent, acceptable, and trustworthy intelligent systems.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"445 - 482"},"PeriodicalIF":12.0,"publicationDate":"2023-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10462-023-10540-1.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43527468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SS-SSAN: a self-supervised subspace attentional network for multi-modal medical image fusion","authors":"Ying Zhang,&nbsp;Rencan Nie,&nbsp;Jinde Cao,&nbsp;Chaozhen Ma,&nbsp;Chengchao Wang","doi":"10.1007/s10462-023-10529-w","DOIUrl":"10.1007/s10462-023-10529-w","url":null,"abstract":"<div><p>Multi-modal medical image fusion (MMIF) is used to merge multiple modes of medical images for better imaging quality and more comprehensive information, such that enhancing the reliability of clinical diagnosis. Since different types of medical images have different imaging mechanisms and focus on different pathological tissues, how to accurately fuse the information from various medical images has become an obstacle in image fusion research. In this paper, we propose a self-supervised subspace attentional framework for multi-modal image fusion, which is constructed by two sub-networks, i.e., the feature extract network and the feature fusion network. We implement a self-supervised strategy that facilitates the framework adaptively extracts the features of source images with the reconstruction of the fused image. Specifically, we adopt a subspace attentional Siamese Weighted Auto-Encoder as a feature extractor to extract the source image features including local and global features at first. Then, the extracted features are given into a weighted fusion decoding network to reconstruct the fused result, and the shallow features from the extractor are used to assist reconstruct the fused image. Finally, the feature extractor adaptively extracts the optimal features according to the fused results by simultaneously training the two sub-networks. Furthermore, to achieve better fusion results, we design a novel weight estimation in the weighted fidelity loss that measures the importance of each pixel by calculating a mixture of salient features and local contrast features of the image. Experiments demonstrate that our method gives the best results compared with other state-of-the-art fusion approaches.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"421 - 443"},"PeriodicalIF":12.0,"publicationDate":"2023-06-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45522570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Brain works principle followed by neural information processing: a review of novel brain theory","authors":"Rubin Wang,&nbsp;Yihong Wang,&nbsp;Xuying Xu,&nbsp;Yuanxi Li,&nbsp;Xiaochuan Pan","doi":"10.1007/s10462-023-10520-5","DOIUrl":"10.1007/s10462-023-10520-5","url":null,"abstract":"<div><p>The way the brain work and its principle of work has long been a big scientific question that scientists have dreamed of solving. However, as is known to all, the brain works at different levels, and the operation at different levels is interactional and mutually coupled. Unfortunately, until now, we still do not know how the nervous system at different levels is interacting and coupling with each other. This review provides some preliminary discussions on how to address these scientific questions, for which we propose a novel theory of the brain called neural energy. Such a theoretical and research approach can couple neural information with neural energy to address the interactions of the nervous system at various levels. Therefore, this review systematically summarizes the neural energy theories and methods proposed by our research in the field of brain science, as well as the internal relationship between mechanics and neural energy theory. Focuses on how to construct a Wang–Zhang (W–Z) neuron model equivalent to Hodgkin–Huxley (H–H) model by using the idea of analytical dynamics. Then, based on this model, we proposed a large-scale neural model and a theoretical framework of global neural coding of the brain in the field of neuroscience. It includes information processing of multiple sensory and perceptual nervous systems such as visual perception, neural mechanism of coupling between default mode network and functional network of brain, memory switching and brain state switching, brain navigation, prediction of new working mechanism of neurons, and interpretation of experimental phenomena that are difficult to be explained by neuroscience. It is proved that the new W–Z neuron model and neural energy theory have unique functions and advantages in neural modeling, neural information processing and methodology. The idea of large-scale neuroscience research with neural energy as the core will provide a potentially powerful research method for promoting the fusion of experimental neuroscience and theoretical neuroscience in the future, and propose a widely accepted brain theory system between experimental neuroscience and theoretical neuroscience. It is of great scientific significance to abandon the shortcomings of reductive and holism research methods in the field of neuroscience, and effectively integrate their respective advantages in methodology.</p></div>","PeriodicalId":8449,"journal":{"name":"Artificial Intelligence Review","volume":"56 1","pages":"285 - 350"},"PeriodicalIF":12.0,"publicationDate":"2023-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10462-023-10520-5.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"45277322","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}