Article
Computer Science, Artificial Intelligence
Pengcheng Wen, Yuhan Zhang, Guihua Wen
Summary: Features of data are critical to classification, but with limited data, obtaining suitable features can be difficult and result in poor performance. This paper proposes a novel approach to automatically learn discriminative features from an irrelevant domain for a given classification task. The method computes central vectors of each class in the irrelevant domain as learning objectives, and uses machine learning to learn features for each sample in the target domain. Unlike transfer learning, this method does not require domain similarity and can learn features from domains with high discrimination. The method is general, simple, efficient, and validated through extensive experiments.
COMPLEX & INTELLIGENT SYSTEMS
(2023)
Article
Environmental Sciences
Han Gao, Jinhui Guo, Peng Guo, Xiuwan Chen
Summary: This work proposes a new object-oriented deep learning framework that leverages residual networks to learn adjacent feature representations for operational land cover mapping tasks. By incorporating geographic object-based image analysis as an auxiliary module, the computational burden of spatial reasoning is reduced and classification boundaries are optimized, leading to better classification accuracy compared to traditional methods.
Article
Engineering, Multidisciplinary
M. Kavitha, R. Gayathri, Kemal Polat, Adi Alhudhaif, Fayadh Alenezi
Summary: This paper introduces an enhanced CNN method for hyperspectral image classification, aiming to improve classification accuracy by merging convolutional layer outputs and using a 1 x 1 convolution layer for feature extraction.
Article
Mathematics
Alma Y. Alanis, Oscar D. Sanchez, Alonso Vaca-Gonzalez, Eduardo Rangel-Heras
Summary: Time series classification is a challenging and exciting problem in data mining. This study presents a classification and diagnosis scheme using deep neural networks for diseases, specifically diabetes mellitus and poor glucose tolerance, in early detection. The results show that deep neural networks have a high accuracy rate, indicating their potential in improving disease diagnosis and classification.
Article
Computer Science, Artificial Intelligence
Musa Peker
Summary: A novel deep learning-based hybrid model called CNN-CVWNN is proposed for hyperspectral image classification, which uses CNN to extract image representation and CVWNN to classify features. Experiments show that the proposed method increases classification accuracy compared to other approaches.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Multidisciplinary Sciences
Zabit Hameed, Begonya Garcia-Zapirain, Jose Javier Aguirre, Mario Arturo Isaza-Ruget
Summary: This paper proposes a deep learning approach for automatic classification of breast cancer microscopy images, achieving good results. The study found that the performance was better on the original dataset, and stain normalization techniques could not surpass the results of the original dataset.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Interdisciplinary Applications
Yetao Yang, Rongkui Tang, Jinglei Wang, Mengjiao Xia
Summary: In this paper, a deep neural network for semantic labeling task is proposed, which iteratively learns deep features in a hierarchical structure and effectively promotes interactions between different hierarchical levels. The use of residual network enhances performance, and dilated k nearest neighbors and multi-scale grouping are introduced to increase the receptive field.
COMPUTERS & GEOSCIENCES
(2021)
Article
Multidisciplinary Sciences
Jun Ogasawara, Satoru Ikenoue, Hiroko Yamamoto, Motoshige Sato, Yoshifumi Kasuga, Yasue Mitsukura, Yuji Ikegaya, Masato Yasui, Mamoru Tanaka, Daigo Ochiai
Summary: The study introduces a new deep neural network model (CTG-net) for detecting compromised fetal status, which successfully classifies and evaluates the performance of CTG data by extracting temporal patterns and interrelationships between fetal heart rate and uterine contraction signals.
SCIENTIFIC REPORTS
(2021)
Article
Computer Science, Artificial Intelligence
Chang Jingfei, Lu Yang, Xue Ping, Wei Xing, Wei Zhen
Summary: The study introduces a novel channel pruning method to reduce storage and calculation costs by directly removing unimportant channels and related convolutional kernels in convolutional layers. After retraining, the compact network outperforms the original network, even with a large pruning amplitude.
JOURNAL OF INTELLIGENT & FUZZY SYSTEMS
(2021)
Article
Computer Science, Information Systems
Qiang Li, Yingjian Yang, Yingwei Guo, Wei Li, Yang Liu, Han Liu, Yan Kang
Summary: Researchers constructed different datasets to test the performance of nine mainstream image classification networks, and found that some of these networks showed better classification performance. The experimental results analyzed the sensitivity of factors that influence the stability of image classification networks.
Article
Biology
Fan Zhang, Yuelei Xu, Zhaoyun Zhou, Han Zhang, Ke Yang
Summary: Preoperative assessment of tracheal intubation difficulty is crucial in anesthesia practice. Current AI methods for Mallampati classification are unreliable, relying solely on doctors' experience. This study proposes a new automatic Mallampati classification method that combines deep features and handcrafted features to improve the accuracy of difficulty assessment in tracheal intubation.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Muna Elsadig, Ashraf Osman Ibrahim, Shakila Basheer, Manal Abdullah Alohali, Sara Alshunaifi, Haya Alqahtani, Nihal Alharbi, Wamda Nagmeldin
Summary: This study introduces a novel URL phishing detection technique based on BERT feature extraction and deep learning. The proposed method achieves a high accuracy rate of 96.66% in detecting phishing URLs, making it efficient and reliable.
Article
Computer Science, Artificial Intelligence
Hua-Nong Ting, Yao-Mun Choo, Azanna Ahmad Kamar
Summary: Previous studies mainly used a single speech feature to classify asphyxia cry among infants. This study investigated the use of multiple speech features and deep learning models to classify asphyxia cry. The results showed that DNN models with multiple hidden layers performed well in classifying normal and asphyxia cry when using hybrid features.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Health Care Sciences & Services
Shih-Yen Hsu, Chi-Yuan Wang, Yi-Kai Kao, Kuo-Ying Liu, Ming-Chia Lin, Li-Ren Yeh, Yi-Ming Wang, Chih- Chen, Feng-Chen Kao
Summary: This study explores the use of a deep neural network algorithm for the classification of mammography images. Based on a retrospective study, imaging data from actual clinical cases were collected and classified using the BI-RADS system. The research results show that the model has high accuracy, sensitivity, and specificity.
Article
Computer Science, Artificial Intelligence
Gao Huang, Zhuang Liu, Geoff Pleiss, Laurens van der Maaten, Kilian Q. Weinberger
Summary: Recent work has shown that adding shorter connections in convolutional networks can make the network deeper, more accurate, and more efficient in training. This paper introduces Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward manner. DenseNets alleviate the vanishing-gradient problem, encourage feature reuse, and improve parameter efficiency, leading to significant improvements in object recognition tasks.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Computer Science, Theory & Methods
Nilesh Chakraborty, Jianqiang Li, Victor C. M. Leung, Samrat Mondal, Yi Pan, Chengwen Luo, Mithun Mukherjee
Summary: This paper presents a comprehensive survey of honeyword-based authentication techniques, covering twenty-three techniques reported since 2013. The paper aims to help readers understand the practical workings of honeyword-based security mechanisms, compare existing techniques, and identify gaps and research opportunities.
ACM COMPUTING SURVEYS
(2023)
Article
Biochemical Research Methods
Chengqian Lu, Lishen Zhang, Min Zeng, Wei Lan, Guihua Duan, Jianxin Wang
Summary: Emerging evidence suggests that circRNAs, with their covalently closed loop structures, can serve as promising biomarkers for diagnosis in pathogenic processes. Computational approaches provide a cost-effective way to identify circRNA-disease associations by aggregating multi-source pathogenesis data and inferring potential associations at the system level. The proposed CDHGNN model, based on edge-weighted graph attention and heterogeneous graph neural networks, outperforms state-of-the-art algorithms in predicting circRNA-disease associations and can identify specific molecular associations and investigate biomolecular regulatory relationships in pathogenesis.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Engineering, Electrical & Electronic
Zhuofan Liao, Yinbao Ma, Jiawei Huang, Jianxin Wang
Summary: This paper proposes an energy-aware 3D-deployment method for Unmanned Aerial Vehicles (UAVs) called 3D-UAV, aiming to ensure a high uplink rate and minimize the number of UAVs in the Internet of Vehicles (IoV) with Highway Interchange. By dividing vehicles into clusters and optimizing the flight altitude of UAVs, this method can cover all vehicles in the bridge scenario and outperform other methods in terms of uplink rate and energy.
IEEE TRANSACTIONS ON COMMUNICATIONS
(2023)
Article
Biochemical Research Methods
Qichang Zhao, Guihua Duan, Mengyun Yang, Zhongjian Cheng, Yaohang Li, Jianxin Wang
Summary: The identification of drug-target relations (DTRs) is crucial in drug development. Traditional methods treating DTRs as drug-target interactions (DTIs) suffer from the lack of reliable negative samples and fail to consider many important aspects of DTRs. With the availability of drug-protein binding affinity data, predicting DTRs as a regression problem of drug-target affinities (DTAs) using deep learning architectures has become feasible. In this study, a deep learning-based model named AttentionDTA is proposed, which utilizes attention mechanism to predict DTAs. The model demonstrates superior performance compared to state-of-the-art methods and successfully extracts protein and drug features to better predict drug-target affinities.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Qichang Zhao, Guihua Duan, Haochen Zhao, Kai Zheng, Yaohang Li, Jianxin Wang
Summary: Drug discovery and drug repurposing benefit from the application of deep learning in predicting drug-target interactions (DTIs). A novel model called GIFDTI is proposed to address the challenges of representing local chemical environments, encoding long-distance relationships, and modeling intermolecular interactions. Evaluation results demonstrate that GIFDTI outperforms state-of-the-art methods in DTI prediction. Case studies also validate the accuracy and cost-effectiveness of the model. The code for GIFDTI is available at https://github.com/zhaoqichang/GIFDTI.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Xiaoqing Peng, Wenjin Zhang, Wanxin Cui, Binrong Ding, Qingtong Lyu, Jianxin Wang
Summary: Alzheimer's disease (AD) is a common neurodegenerative disease, and DNA methylation is closely related to its pathological mechanism. A database named ADmeth has been designed to collect AD-related differential methylation data, containing 16,709 items identified from various brain regions and cell types in the blood, including 209 genes, 2,229 regions, and 14,271 CpG sites. The ADmeth database provides user-friendly functions for searching, submitting, and downloading data, aiming to facilitate research on the pathological mechanism of AD and non-invasive diagnosis using cell-free DNA.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2023)
Article
Computer Science, Hardware & Architecture
Jiawei Huang, Wenlu Zhang, Yijun Li, Lin Li, Zhaoyi Li, Jin Ye, Jianxin Wang
Summary: Identifying heavy flows is crucial for network management, but it is challenging to detect heavy flow quickly and accurately in highly dynamic traffic and rapidly growing network capacity. Existing schemes trade-off efficiency, accuracy, and speed, requiring large memory for acceptable performance. To address this, ChainSketch is proposed, with advantages in memory efficiency, accuracy, and fast detection. ChainSketch utilizes selective replacement, hash chain, and compact structure, significantly improving F1-score compared to existing solutions, especially in small memory conditions.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Article
Computer Science, Hardware & Architecture
Jingling Liu, Jiawei Huang, Weihe Li, Jianxin Wang, Tian He
Summary: Datacenter networks often face path asymmetry, leading to problems like packet reordering and under-utilization of multiple paths. In this paper, we propose a load balancing mechanism called AG that adaptively adjusts switching granularity based on the degree of topology asymmetry. We also design a switch-based scheme to measure the difference of one-way delay, allowing accurate detection of topology asymmetry. Experimental results show that AG outperforms existing load balancing schemes in terms of flow completion time.
IEEE-ACM TRANSACTIONS ON NETWORKING
(2023)
Article
Biochemical Research Methods
Haochen Zhao, Xiaoyu Zhang, Qichang Zhao, Yaohang Li, Jianxin Wang
Summary: In this paper, a deep learning framework called MSDRP is proposed for drug response prediction. MSDRP captures interactions between drugs and cell lines using an interaction module, and integrates multiple associations/interactions between drugs and biological entities through similarity network fusion algorithms. Experimental results demonstrate the excellent performance of our model in all performance measures for all experiments.
Article
Biochemical Research Methods
Bin Wang, Kun Liu, Yaohang Li, Jianxin Wang
Summary: In this research, a novel method called DFHiC is proposed to generate high-resolution Hi-C matrix from low-resolution Hi-C matrix using the dilated convolutional neural network framework. DFHiC can reliably and accurately improve the resolution of Hi-C matrix, and the super-resolution Hi-C data enhanced by DFHiC is more similar to real high-resolution Hi-C data in terms of both chromatin significant interactions and identifying topologically associating domains.
Article
Computer Science, Artificial Intelligence
Wei Lan, Tianchuan Yang, Qingfeng Chen, Shichao Zhang, Yi Dong, Huiyu Zhou, Yi Pan
Summary: This article proposes a novel multiview subspace clustering method, named LSGMC, to address the problems of ignoring consistent information and angular information in existing methods. LSGMC pursues a consistent low-rank structure across views and guarantees weight consistency using a symmetry constraint. It captures the inherent structure of data by utilizing fusion mechanism and employs the Schatten p-norm to obtain a low-rank coefficient matrix. Experimental results on 11 datasets demonstrate the superiority of LSGMC in clustering performance compared with ten state-of-the-art multiview clustering methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Biochemical Research Methods
Neng Huang, Minghua Xu, Fan Nie, Peng Ni, Chuan-Le Xiao, Feng Luo, Jianxin Wang
Summary: We developed a deep learning-based method called NanoSNP for identifying SNP sites in low-coverage Nanopore sequencing data. NanoSNP uses a multi-step, multi-scale, and haplotype-aware pipeline to detect SNP sites and predict genotypes. Comparison with other methods showed that NanoSNP outperformed Clair, Pepper-DeepVariant, and NanoCaller in identifying SNPs, especially in difficult-to-map regions and the major histocompatibility complex regions of the human genome. NanoSNP performed comparably to Clair3 when coverage exceeded 16x.
Article
Biochemistry & Molecular Biology
Hao Wu, Jovial Niyogisubizo, Keliang Zhao, Jintao Meng, Wenhui Xi, Hongchang Li, Yi Pan, Yanjie Wei
Summary: The automatic detection of cells in microscopy image sequences is a significant task in biomedical research. We propose a novel weakly supervised learning cell detection and tracking framework that uses incomplete initial labels to train the deep neural network. Our method has been evaluated and proven to be effective in experiments.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Article
Computer Science, Information Systems
Weihe Li, Jiawei Huang, Wenjun Lyu, Baoshen Guo, Wanchun Jiang, Jianxin Wang
Summary: Current ABR algorithms do not pay enough attention to audio bitrate selection, assuming it has minimal impact on video selection. However, with the advancement of audio technologies, audio bitrate can significantly affect video selection and viewing experience. To address this issue, we propose a deep reinforcement learning-based ABR algorithm that considers both audio and video quality, achieving significant improvements in overall viewing quality.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Information Systems
Shigeng Zhang, Zijing Ma, Kaixuan Lu, Xuan Liu, Jia Liu, Song Guo, Albert Y. Zomaya, Jian Zhang, Jianxin Wang
Summary: This paper presents HearMe, an accurate and real-time lip-reading system built on commercial RFID devices. HearMe can help people with speech disorders communicate and interact with the world effectively. By utilizing effective data collection, signal pattern extraction, and feature extraction techniques, HearMe achieves high accuracy in mouth motion recognition and word-level recognition. The use of transfer learning enhances the model's robustness in different environments.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Artificial Intelligence
Rui Lv, Dingheng Wang, Jiangbin Zheng, Zhao-Xu Yang
Summary: In this paper, the authors investigate tensor decomposition for neural network compression. They analyze the convergence and precision of tensor mapping theory, validate the rationality of tensor mapping and its superiority over traditional tensor approximation based on the Lottery Ticket Hypothesis. They propose an efficient method called 3D-KCPNet to compress 3D convolutional neural networks using the Kronecker canonical polyadic (KCP) tensor decomposition. Experimental results show that 3D-KCPNet achieves higher accuracy compared to the original baseline model and the corresponding tensor approximation model.
Article
Computer Science, Artificial Intelligence
Xiangkun He, Zhongxu Hu, Haohan Yang, Chen Lv
Summary: In this paper, a novel constrained multi-objective reinforcement learning algorithm is proposed for personalized end-to-end robotic control with continuous actions. The approach trains a single model using constraint design and a comprehensive index to achieve optimal policies based on user-specified preferences.
Article
Computer Science, Artificial Intelligence
Zhijian Zhuo, Bilian Chen, Shenbao Yu, Langcai Cao
Summary: In this paper, a novel method called Expansion with Contraction Method for Overlapping Community Detection (ECOCD) is proposed, which utilizes non-negative matrix factorization to obtain disjoint communities and applies expansion and contraction processes to adjust the degree of overlap. ECOCD is applicable to various networks with different properties and achieves high-quality overlapping community detection.
Article
Computer Science, Artificial Intelligence
Yizhe Zhu, Chunhui Zhang, Jialin Gao, Xin Sun, Zihan Rui, Xi Zhou
Summary: In this work, the authors propose a Contrastive Spatio-Temporal Distilling (CSTD) approach to improve the detection of high-compressed deepfake videos. The approach leverages spatial-frequency cues and temporal-contrastive alignment to fully exploit spatiotemporal inconsistency information.
Review
Computer Science, Artificial Intelligence
Laijin Meng, Xinghao Jiang, Tanfeng Sun
Summary: This paper provides a review of coverless steganographic algorithms, including the development process, known contributions, and general issues in image and video algorithms. It also discusses the security of coverless steganography from theoretical analysis to actual investigation for the first time.
Article
Computer Science, Artificial Intelligence
Yajie Bao, Tianwei Xing, Xun Chen
Summary: Visual question answering requires processing multi-modal information and effective reasoning. Neural-symbolic learning is a promising method, but current approaches lack uncertainty handling and can only provide a single answer. To address this, we propose a confidence based neural-symbolic approach that evaluates NN inferences and conducts reasoning based on confidence.
Article
Computer Science, Artificial Intelligence
Anh H. Vo, Bao T. Nguyen
Summary: Interior style classification is an interesting problem with potential applications in both commercial and academic domains. This project proposes a method named ISC-DeIT, which combines data-efficient image transformer architectures and knowledge distillation, to address the interior style classification problem. Experimental results demonstrate a significant improvement in predictive accuracy compared to other state-of-the-art methods.
Article
Computer Science, Artificial Intelligence
Shashank Kotyan, Danilo Vasconcellos Vargas
Summary: This article introduces a novel augmentation technique called Dynamic Scanning Augmentation to improve the accuracy and robustness of Vision Transformer (ViT). The technique leverages dynamic input sequences to adaptively focus on different patches, resulting in significant changes in ViT's attention mechanism. Experimental results demonstrate that Dynamic Scanning Augmentation outperforms ViT in terms of both robustness to adversarial attacks and accuracy against natural images.
Article
Computer Science, Artificial Intelligence
Hiba Alqasir, Damien Muselet, Christophe Ducottet
Summary: The article proposes a solution to improve the learning process of a classification network by providing shape priors, reducing the need for annotated data. The solution is tested on cross-domain digit classification tasks and a video surveillance application.
Article
Computer Science, Artificial Intelligence
Dexiu Ma, Mei Liu, Mingsheng Shang
Summary: This paper proposes a method using neural dynamics solvers to solve infinity-norm optimization problems. Two improved solvers are constructed and their effectiveness and superiority are demonstrated through theoretical analysis and simulation experiments.
Article
Computer Science, Artificial Intelligence
Francesco Gregoretti, Giovanni Pezzulo, Domenico Maisto
Summary: Active Inference is a computational framework that uses probabilistic inference and variational free energy minimization to describe perception, planning, and action. cpp-AIF is a header-only C++ library that provides a powerful tool for implementing Active Inference for Partially Observable Markov Decision Processes through multi-core computing. It is cross-platform and improves performance, memory management, and usability compared to existing software.
Article
Computer Science, Artificial Intelligence
Zelin Ying, Dawei Cheng, Cen Chen, Xiang Li, Peng Zhu, Yifeng Luo, Yuqi Liang
Summary: This paper proposes a novel stock market trends prediction framework called SMART, which includes a self-supervised stock technical data sequence embedding model S3E. By training with multiple self-supervised auxiliary tasks, the model encodes stock technical data sequences into embeddings and uses the learned sequence embeddings for predicting stock market trends. Extensive experiments on China A-Shares market and NASDAQ market prove the high effectiveness of our model in stock market trends prediction, and its effectiveness is further validated in real-world applications in a leading financial service provider in China.
Article
Computer Science, Artificial Intelligence
Hao Li, Hao Jiang, Dongsheng Ye, Qiang Wang, Liang Du, Yuanyuan Zeng, Liu Yuan, Yingxue Wang, C. Chen
Summary: DHGAT1, a dynamic hyperbolic graph attention network, utilizes hyperbolic metric properties to embed dynamic graphs. It employs a spatiotemporal self-attention mechanism and weighted node representations, resulting in excellent performance in link prediction tasks.
Article
Computer Science, Artificial Intelligence
Jiehui Huang, Zhenchao Tang, Xuedong He, Jun Zhou, Defeng Zhou, Calvin Yu-Chian Chen
Summary: This study proposes a progressive learning multi-scale feature blending model for image deraining tasks. The model utilizes detail dilation and texture extraction to improve the restoration of rainy images. Experimental results show that the model achieves near state-of-the-art performance in rain removal tasks and exhibits better rain removal realism.
Article
Computer Science, Artificial Intelligence
Lizhi Liu, Zilin Gao, Yinhe Wang, Yongfu Li
Summary: This paper proposes a novel discrete-time interconnected model for depicting complex dynamical networks. The model consists of nodes and edges subsystems, which consider the dynamic characteristic of both nodes and edges. By designing control strategies and coupling modes, the stabilization and synchronization of the network are achieved. Simulation results demonstrate the effectiveness of the proposed methods.