检索结果-内蒙古大学图书馆

UniEmoX: Cross-modal Semantic-Guided Large-Scale Pretraining for Universal Scene Emotion Perception

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Chuang Sun, Xiao Liu, Zhi School of Artificial Intelligence Anhui University Hefei230601 China Institute of Artificial Intelligence Hefei Comprehensive National Science Center Hefei230088 China Anhui Province Key Laboratory of Affective Computing and Advanced Intelligent Machines School of Computer Science and Information Engineering Hefei University of Technology Hefei230009 China Department of Computer and Network Engineering The University of Electro-Communications Chofu-shi Tokyo182-8585 Japan

Visual emotion analysis holds significant research value in both computer vision and psychology. However, existing methods for visual emotion analysis suffer from limited generalizability due to the ambiguity of emotion perception and the diversity of data scenarios. To tackle this issue, we introduce UniEmoX, a cross-modal semantic-guided large-scale pretraining framework. Inspired by psychological research emphasizing the inseparability of the emotional exploration process from the interaction between individuals and their environment, UniEmoX integrates scene-centric and person-centric low-level image spatial structural information, aiming to derive more nuanced and discriminative emotional representations. By exploiting the similarity between paired and unpaired image-text samples, UniEmoX distills rich semantic knowledge from the CLIP model to enhance emotional embedding representations more effectively. To the best of our knowledge, this is the first large-scale pretraining framework that integrates psychological theories with contemporary contrastive learning and masked image modeling techniques for emotion analysis across diverse scenarios. Additionally, we develop a visual emotional dataset titled Emo8. Emo8 samples cover a range of domains, including cartoon, natural, realistic, science fiction and advertising cover styles, covering nearly all common emotional scenes. Comprehensive experiments conducted on six benchmark datasets across two downstream tasks validate the effectiveness of UniEmoX. The source code is available at https://***/chincharles/u-emo. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

GraphWGAN: Graph Representation Learning with Wasserstein Generative Adversarial networks

学校读者我要写书评

暂无评论

GraphWGAN: Graph Representation Learning with Wasserstein Ge...

International Conference on Big data and Smart computing (BIGCOMP)

作者： Rong Yan Huawei Shen Cao Qi Keting Cen Li Wang College of Big Data Taiyuan University Of Technology Taiyuan China CAS Key Lab of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China Institute of Computing Technology Beijing China

ISBN: (数字)9781728160344

ISBN: (纸本)9781728160351

Graph representation learning aims to represent vertices as low-dimensional and real-valued vectors to facilitate subsequent downstream tasks, i.e., node classification, link predictions. Recently, some novel graph representation learning frameworks, which try to approximate the underlying true connectivity distribution of the vertices, show their superiority. These methods characterize the distance between the true connectivity distribution and generated connectivity distribution by Kullback-Leibler or Jensen-Shannon divergence. However, since these divergences are not continuous with respect to the generator's parameters, such methods easily lead to unstable training and poor convergence. In contrast, Wasserstein distance is continuous and differentiable almost everywhere, which means it can produce more reliable gradient, allowing the training more stable and more convergent. In this paper, we utilize Wasserstein distance to characterize the distance between the underlying true connectivity distribution and generated distribution in graph representation learning. Experimental results show that the accuracy of our method exceeds existing baselines in tasks of both node classification and link prediction.

关键词： Hafnium Conferences Big data

Channel Drop Out:A Simple Way to Prevent CNN from Overfitting in Motor Imagery Based BCI

学校读者我要写书评

暂无评论

Channel Drop Out:A Simple Way to Prevent CNN from Overfittin...

2021国际计算机前沿大会

作者： Jing Luo Yaojie Wang Rong Xu Guangming Liu Xiaofan Wang Yijing Gong Shaanxi Key Laboratory for Network Computing and Security Technology School of Computer Science and EngineeringXi'an University of Technology Shaanxi Province Institute of Water Resources and Electric Power Investigation and Design

With the development of deep learning,many motor imagery brain-computer interfaces based on convolutional neural networks(CNNs) show outstanding ***,the trial number of EEG in the training set is usually limited,and redundancy extensively exists in multiple channel ***,overfitting often appears in CNN based motor imagery recognition model and greatly affects the performances of *** this paper,channel drop out is proposed to address this problem by data augmentation and ensemble ***,one of all EEG channels will be dropped and replaced by the mean signal of all EEG *** this way,the trial number in the training set was enlarged by channel drop *** at the testing stage,all the EEG trials processed by channel drop out were fed to the CNN model and the average output probabilities of them were applied to determine the *** experiments were conducted on two popular CNN models applied in motor imagery BCI and BCI Competition IV datasets 2 a to verify the performances of the proposed channel drop out *** results show that average improvements provided by channel drop out in two-category or four-category motor imagery classification are 2.83%and 2.65% compared with the original CNN *** the channel drop out approach significantly improves the performances of motor imagery based BCI.

关键词：

Graph Reciprocal Neural networks by Abstracting Node as Attribute

学校读者我要写书评

暂无评论

Graph Reciprocal Neural Networks by Abstracting Node as Attr...

IEEE International Conference on data Mining (ICDM)

作者： Liang Yang Jiayi Wang Dongxiao He Chuan Wang Xiaochun Cao Bingxin Niu Zhen Wang School of Artificial Intelligence Hebei University of Technology Tianjin China College of Intelligence and Computing Tianjin University Tianjin China State Key Laboratory of Information Security Institute of Information Engineering CAS Beijing China School of Cyber Science and Technology Shenzhen Campus Sun Yat-sen University Shenzhen China School of Artificial Intelligence OPtics and ElectroNics (iOPEN) School of Cybersecurity Northwestern Polytechnical University Xi’an China

Graph neural network (GNN) can be formulated as the multiplication of the topology-related matrix (adjacency or Laplacian matrix) and node attribute matrix, i.e., operation in node-wise. Unfortunately, this unified formula reveals two inherent drawbacks. Firstly, the topology and node attribute are not reciprocal but biased. From employment, the topology information is repeatedly employed, while the node attribute is only used once. From parameterization perspective, the node attribute is parameterized with highly expressive MLPs, while topology is not. Secondly, the graph topology can not be fully explored. Only the local pairwise relation is explored, but the mesoscopic community structure, which is one of the most prominent characteristics of networks, is ignored. To alleviate these issues, this paper proposes the Graph Reciprocal network (GRN) by treating node attribute and topology reciprocal. Firstly, it is illustrated that the node can be regarded and utilized as another kind of attribute. Secondly, a novel node representation scheme is proposed from the theory of Quadratic networks, with a theoretical guarantee of the fine-grained element-wise product of the representations of the topology and attribute. Extensive experiments demonstrate the superior performance and robustness of the proposed GRN.

关键词：

A Node-collaboration-informed Graph Convolutional network for Precise Representation to Undirected Weighted Graphs

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wang, Ying Yuan, Ye Luo, Xin The School of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing400065 China The Chongqing Key Laboratory of Big Data and Intelligent Computing Chongqing Engineering Research Center of Big Data Application for Smart Cities Chongqing Institute of Green and Intelligent Technology Chinese Academy of Sciences Chongqing400714 China The College of Computer and Information Science Southwest University Chongqing400715 China

An undirected weighted graph (UWG) is frequently adopted to describe the interactions among a solo set of nodes from real applications, such as the user contact frequency from a social network services system. A graph convolutional network (GCN) is widely adopted to perform representation learning to a UWG for subsequent pattern analysis tasks such as clustering or missing data estimation. However, existing GCNs mostly neglects the latent collaborative information hidden in its connected node pairs. To address this issue, this study proposes to model the node collaborations via a symmetric latent factor analysis model, and then regards it as a node-collaboration module for supplementing the collaboration loss in a GCN. Based on this idea, a Node-collaboration-informed Graph Convolutional network (NGCN) is proposed with three-fold ideas: a) Learning latent collaborative information from the interaction of node pairs via a node-collaboration module;b) Building the residual connection and weighted representation propagation to obtain high representation capacity;and c) Implementing the model optimization in an end-to-end fashion to achieve precise representation to the target UWG. Empirical studies on UWGs emerging from real applications demonstrate that owing to its efficient incorporation of node-collaborations, the proposed NGCN significantly outperforms state-of-the-art GCNs in addressing the task of missing weight estimation. Meanwhile, its good scalability ensures its compatibility with more advanced GCN extensions, which will be further investigated in our future studies. Copyright © 2022, The Authors. All rights reserved.

关键词： Convolution

Deep latent low-rank fusion network for progressive subspace discovery 29

学校读者我要写书评

暂无评论

Deep latent low-rank fusion network for progressive subspace...

29th International Joint Conference on Artificial Intelligence, IJCAI 2020

作者： Zhang, Zhao Ren, Jiahuan Zhang, Zheng Liu, Guangcan Key Laboratory of Knowledge Engineering with Big Data Hefei University of Technology China School of Computer Science and Technology Soochow University China Bio-Computing Research Center Harbin Institute of Technology Shenzhen China School of Information and Control Nanjing University of Information Science and Technology China

ISBN: (纸本)9780999241165

Low-rank representation is powerful for recovering and clustering the subspace structures, but it cannot obtain deep hierarchical information due to the single-layer mode. In this paper, we present a new and effective strategy to extend the single-layer latent low-rank models into multiple-layers, and propose a new and progressive Deep Latent Low-Rank Fusion network (DLRF-Net) to uncover deep features and structures embedded in input data. The basic idea of DLRF-Net is to refine features progressively from the previous layers by fusing the subspaces in each layer, which can potentially obtain accurate features and subspaces for representation. To learn deep information, DLRF-Net inputs shallow features of the last layers into subsequent layers. Then, it recovers the deeper features and hierarchical information by congregating the projective subspaces and clustering subspaces respectively in each layer. Thus, one can learn hierarchical subspaces, remove noise and discover the underlying clean subspaces. Note that most existing latent low-rank coding models can be extended to multilayers using DLRF-Net. Extensive results show that our network can deliver enhanced performance over other related frameworks. © 2020 Inst. Sci. inf., Univ. Defence in Belgrade. All rights reserved.

关键词：

Cross-Domain Dual-Functional OFDM Waveform Design for Accurate Sensing/Positioning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhang, Fan Mao, Tianqi Liu, Ruiqi Han, Zhu Chen, Sheng Wang, Zhaocheng Beijing National Research Center for Information Science and Technology Department of Electronic Engineering Tsinghua University Beijing100084 China Tsinghua Shenzhen International Graduate School Shenzhen518055 China The State Key Laboratory of CNS/ATM Beijing Institute of Technology Beijing100081 China The MIIT Key Laboratory of Complex-field Intelligent Sensing Beijing Institute of Technology Beijing100081 China The Wireless and Computing Research Institute ZTE Corporation Beijing100029 China The State Key Laboratory of Mobile Network and Mobile Multimedia Technology Shenzhen518055 China The Department of Electrical and Computer Engineering University of Houston HoustonTX77004 United States The Department of Computer Science and Engineering Kyung Hee University Seoul446-701 Korea Republic of The School of Electronics and Computer Science University of Southampton SouthamptonSO17 1BJ United Kingdom

Orthogonal frequency division multiplexing (OFDM) has been widely recognized as the representative waveform for 5G wireless networks, which can directly support sensing/positioning with existing infrastructure. To guarantee superior sensing/positioning accuracy while supporting high-speed communications simultaneously, the dual functions tend to be assigned with different resource elements (REs) due to their diverse design requirements. This motivates optimization of resource allocation/waveform design across time, frequency, power and delay-Doppler domains. Therefore, this article proposes two cross-domain waveform optimization strategies for effective convergence of OFDM-based communications and sensing/positioning, following communication- and sensing-centric criteria, respectively. For the communication-centric design, to maximize the achievable data rate, a fraction of REs are optimally allocated for communications according to prior knowledge of the communication channel. The remaining REs are then employed for sensing/positioning, where the sidelobe level and peak-to-average power ratio are suppressed by optimizing its power-frequency and phase-frequency characteristics for sensing performance improvement. For the sensing-centric design, a 'locally' perfect auto-correlation property is ensured for accurate sensing and positioning by adjusting the unit cells of the ambiguity function within its region of interest (RoI). Afterwards, the irrelevant cells beyond RoI, which can readily determine the sensing power allocation, are optimized with the communication power allocation to enhance the achievable data rate. Numerical results demonstrate the superiority of the proposed waveform designs. Copyright © 2023, The Authors. All rights reserved.

关键词： Orthogonal frequency division multiplexing

UOD: Universal One-shot Detection of Anatomical Landmarks

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhu, Heqin Quan, Quan Yao, Qingsong Liu, Zaiyi Zhou, S. Kevin School of Biomedical Engineering Division of Life Sciences and Medicine University of Science and Technology of China Anhui Hefei230026 China Suzhou Institute for Advanced Research University of Science and Technology of China Jiangsu Suzhou215123 China Institute of Computing Technology CAS Beijing100190 China Department of Radiology Guangdong Provincial People's Hospital Guangdong Academy of Medical Sciences Guangzhou China Guangdong Provincial Key Laboratory of Artificial Intelligence in Medical Image Analysis and Application Guangdong Provincial People's Hospital Guangdong Academy of Medical Sciences Guangzhou China

One-shot medical landmark detection gains much attention and achieves great success for its label-efficient training process. However, existing one-shot learning methods are highly specialized in a single domain and suffer domain preference heavily in the situation of multi-domain unlabeled data. Moreover, one-shot learning is not robust that it faces performance drop when annotating a sub-optimal image. To tackle these issues, we resort to developing a domain-adaptive one-shot landmark detection framework for handling multi-domain medical images, named Universal One-shot Detection (UOD). UOD consists of two stages and two corresponding universal models which are designed as combinations of domain-specific modules and domain-shared modules. In the first stage, a domain-adaptive convolution model is self-supervised learned to generate pseudo landmark labels. In the second stage, we design a domain-adaptive transformer to eliminate domain preference and build the global context for multi-domain data. Even though only one annotated sample from each domain is available for training, the domain-shared modules help UOD aggregate all one-shot samples to detect more robust and accurate landmarks. We investigated both qualitatively and quantitatively the proposed UOD on three widely-used public X-ray datasets in different anatomical domains (i.e., head, hand, chest) and obtained state-of-the-art performances in each domain. The code is at https://***/heqin-zhu/UOD_universal_oneshot_detection. © 2023, CC BY-NC-SA.

关键词： Learning systems

PreZ-DGGAN: A Drug Graph GAN Based on Pre-Learning of Implicit Variables 2nd

学校读者我要写书评

暂无评论

PreZ-DGGAN: A Drug Graph GAN Based on Pre-Learning of Implic...

2nd International Conference on Applied Intelligence, ICAI 2024

作者： Liu, Yixin Fan, Yueqin Li, Zhipeng Zhang, Qinhu Big Data and Intelligent Computing Research Center Guangxi Academy of Science Nanning530007 China School of Mechanical Engineering Guangxi University Nanning530004 China Ningbo Institute of Digital Twin Eastern Institute of Technology Ningbo315201 China Institute for Regenerative Medicine Medical Innovation Center and State Key Laboratory of Cardiology School of Medicine Shanghai East Hospital Tongji University Shanghai200123 China College of Advanced Agricultural Sciences Zhejiang Agriculture and Forestry University Hangzhou311300 China

ISBN: (纸本)9789819619061

In the field of drug discovery and development, deep learning techniques have become a powerful tool to accelerate the discovery and development of new drugs. In the design and optimization of lead molecules, generative adversarial network and graph neural network are one of the prime techniques for accelerating research and development, because the former is based on the idea of two-player game to learn drug features to generate effective molecules, while the unstructured inputs of the latter are well-suited for the analysis of molecular data. Therefore, this study develops a new graph generative adversarial model, called PreZ-DGGAN, by inserting the graph neural network layer into the generative adversarial network. Within PreZ-DGGAN, the latent variable input of the generator is pre-learned by an improved graph variational autoencoder, which effectively avoids the mode collapse of the generative adversarial network, whereas the hidden layer of the discriminator is graph neural network layer, which is able to better learn the real molecules and judge the generated molecules as true or false. Experiments on two molecule datasets show that the molecules generated by PreZ-DGGAN have higher validity (100% on average) and uniqueness (91.5% on average), and the generated molecules are more diverse in structure. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Generative adversarial networks