检索结果-内蒙古大学图书馆

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Lu, Yao Chen, Zhiyi Chen, Zehui Hu, Jie Cao, Liujuan Zhang, Shengchuan Xiamen University Media Analytics and Computing Lab China Ustc China

ISBN: (纸本)9781728163277

Instance segmentation has been dominated by the paradigm that predicts masks using local RoI features and simplicity frameworks based on global mask prediction. Despite the comparable performance between local-based and global-based approaches, the AP results of objects on different scales vary significantly. In this paper, we first point out that the key factor to bridging such a gap lies in the utilization of local RoI information for global mask prediction. Then, we observe a 'class-agnostic segmentation' problem exists in the nearby region of interesting objects after implementing the above combination. To overcome this issue, we further propose a CAtegory-kerNelized DYnamic (CANDY) convolution. Benefiting from it, the discriminative ability of the resulting instance segmentation framework, i.e., CANDY-Mask, on foreground objects is significantly enhanced. Extensive experiments on the MS-COCO dataset are conducted to verify the performance of CANDY-Mask. Our proposed CANDY-Mask obtains 48.1% boxes AP and 40.7% masks AP on the MS-COCO test-dev set with ResNet50 backbone, achieving state-of-the-art among various models. © 2023 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Self-Paced Partial Domain-Aware Learning for Face Anti-Spoofing 48

Self-Paced Partial Domain-Aware Learning for Face Anti-Spoof...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Chen, Zhiyi Lu, Yao Deng, Xinzhe Meng, Jia Zhang, Shengchuan Cao, Liujuan Xiamen University Media Analytics and Computing Lab Department of Artificial Intelligence School of Informatics China Youtu Lab Tencent China

ISBN: (纸本)9781728163277

With the widespread deployment of face authentication systems, domain generalization (DG) based face anti-spoofing (FAS) security approaches have drawn growing attention. Existing generalization-based methods always attempt to extract domain-invariant task information from data and eliminate domain-dependent information from representation space. However, they neglect that domain-related information may also contain helpful features for the classification task. To address this issue, we propose a self-paced partial domain-aware framework (SPDA) to preserve domain-related features helpful for the discrimination of fake and real faces, thereby increasing generalization for unseen domains. Specifically, a training strategy based on contrastive learning is adopted to construct domain-adapted and domain-aware task-related representation spaces. Then, a partial domain-aware adaptation module (PDA) is proposed to preserve valuable domain-related information for the task features that the network considers useful for mixture-domain classification. In addition, the proposed self-paced method(SCM) continuously explores potential clusters with insufficient representation to enhance further the feature extractor's capability and the effectiveness of the PDA module. Extensive experiments demonstrate the effectiveness of our method compared to SOTA algorithms. © 2023 IEEE.

关键词： Contrastive Learning Domain Generalization Face Anti-spoofing

来源：评论

学校读者我要写书评

暂无评论

Trust Your Partner's Friends: Hierarchical Cross-Modal Contrastive Pre-Training for Video-Text Retrieval 48

Trust Your Partner's Friends: Hierarchical Cross-Modal Contr...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Xiang, Yuhan Liu, Kaijian Tang, Shixiang Bai, Lei Zhu, Feng Zhao, Rui Lin, Xianming School of Infomatics Xiamen University Media Analytics and Computing Lab Department of Artificial Intelligence 361005 China SenseTime Group Limited China The University of Sydney Australia

ISBN: (纸本)9781728163277

Video-text retrieval has greatly benefited from the massive web video in recent years, while the performance is still limited to the weak supervision from the uncurated data. In this work, we propose to leverage the well-represented information of each original modality and exploit complementary information in two views of the same video, i.e., video clips and captions, by using one view to obtain positive samples with the neighboring samples of the other. Respecting the hierarchical organization of real-world data, we further design a hierarchical cross-modal pre-training method (HCP) to learn good representations in the common embedding space. We evaluate the pre-trained model on three downstream tasks, i.e. text-to-video retrieval, action step localization and video question answering and our method outperforms previous works under the same setting. © 2023 IEEE.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

CANDY: Category-Kernelized Dynamic Convolution for Instance Segmentation

CANDY: Category-Kernelized Dynamic Convolution for Instance ...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Yao Lu Zhiyi Chen Zehui Chen Jie Hu Liujuan Cao Shengchuan Zhang Media Analytics and Computing Lab Xiamen University China USTC China

Instance segmentation has been dominated by the paradigm that predicts masks using local RoI features and simplicity frameworks based on global mask prediction. Despite the comparable performance between local-based and global-based approaches, the AP results of objects on different scales vary significantly. In this paper, we first point out that the key factor to bridging such a gap lies in the utilization of local RoI information for global mask prediction. Then, we observe a ’class-agnostic segmentation’ problem exists in the nearby region of interesting objects after implementing the above combination. To overcome this issue, we further propose a CAtegory-kerNelized DYnamic (CANDY) convolution. Benefiting from it, the discriminative ability of the resulting instance segmentation framework, i.e., CANDY-Mask, on foreground objects is significantly enhanced. Extensive experiments on the MS-COCO dataset are conducted to verify the performance of CANDY-Mask. Our proposed CANDY-Mask obtains 48.1% boxes AP and 40.7% masks AP on the MS-COCO test-dev set with ResNet50 backbone, achieving state-of-the-art among various models.

关键词： Visualization Convolution Detectors Benchmark testing Acoustics Speech processing Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

Towards End-to-end Semi-supervised Learning for One-stage Object Detection

arXiv

引用

arXiv 2023年

作者： Luo, Gen Zhou, Yiyi Jin, Lei Sun, Xiaoshuai Ji, Rongrong The Media Analytics and Computing Lab Department of Artificial Intelligence Xiamen University China

Semi-supervised object detection (SSOD) is a research hot spot in computer vision, which can greatly reduce the requirement for expensive bounding-box annotations. Despite great success, existing progress mainly focuses on two-stage detection networks like FasterRCNN, while the research on one-stage detectors is often ignored. In this paper, we focus on the semi-supervised learning for the advanced and popular one-stage detection network YOLOv5. Compared with Faster-RCNN, the implementation of YOLOv5 is much more complex, and the various training techniques used in YOLOv5 can also reduce the benefit of SSOD. In addition to this challenge, we also reveal two key issues in one-stage SSOD, which are low-quality pseudo-labeling and multi-task optimization conflict, respectively. To address these issues, we propose a novel teacher-student learning recipe called OneTeacher with two innovative designs, namely Multi-view Pseudo-label Refinement (MPR) and Decoupled Semi-supervised Optimization (DSO). In particular, MPR improves the quality of pseudo-labels via augmented-view refinement and global-view filtering, and DSO handles the joint optimization conflicts via structure tweaks and task-specific pseudo-labeling. In addition, we also carefully revise the implementation of YOLOv5 to maximize the benefits of SSOD, which is also shared with the existing SSOD methods for fair comparison. To validate OneTeacher, we conduct extensive experiments on COCO and Pascal VOC. The extensive experiments show that OneTeacher can not only achieve superior performance than the compared methods, e.g., 15.0% relative AP gains over Unbiased Teacher, but also well handle the key issues in one-stage SSOD. Code is available at: https://***/luogen1996/OneTeacher. © 2023, CC BY.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

SiMaN: Sign-to-Magnitude Network Binarization

引用

IEEE Transactions on Pattern Analysis and Machine Intelligence 2023年第5期45卷 6277-6288页

作者： Lin, Mingbao Ji, Rongrong Xu, Zihan Zhang, Baochang Chao, Fei Lin, Chia-Wen Shao, Ling Xiamen University School of Informatics Media Analytics and Computing Laboratory Department of Artificial Intelligence Xiamen361005 China Tencent Youtu Lab Shanghai200233 China Zhongguancun Lab Beijing100190 China National Tsing Hua University Department of Electrical Engineering The Institute of Communications Engineering Hsinchu30013 Taiwan Terminus Group Beijing100026 China

Binary neural networks (BNNs) have attracted broad research interest due to their efficient storage and computational ability. Nevertheless, a significant challenge of BNNs lies in handling discrete constraints while ensuring bit entropy maximization, which typically makes their weight optimization very difficult. Existing methods relax the learning using the sign function, which simply encodes positive weights into +1s, and -1s otherwise. Alternatively, we formulate an angle alignment objective to constrain the weight binarization to {0,+1} to solve the challenge. In this article, we show that our weight binarization provides an analytical solution by encoding high-magnitude weights into +1s, and 0s otherwise. Therefore, a high-quality discrete solution is established in a computationally efficient manner without the sign function. We prove that the learned weights of binarized networks roughly follow a Laplacian distribution that does not allow entropy maximization, and further demonstrate that it can be effectively solved by simply removing the 2 regularization during network training. Our method, dubbed sign-to-magnitude network binarization (SiMaN), is evaluated on CIFAR-10 and ImageNet, demonstrating its superiority over the sign-based state-of-the-arts. Our source code, experimental settings, training logs and binary models are available at https://***/lmbxμSiMaN. © 1979-2012 IEEE.

关键词： Quantization (signal)

来源：评论

学校读者我要写书评

暂无评论

Deepwalk-aware graph convolutional networks

引用

Science China(Information Sciences) 2022年第5期65卷 81-95页

作者： Taisong JIN Huaqiang DAI Liujuan CAO Baochang ZHANG Feiyue HUANG Yue GAO Rongrong JI Media Analytics and Computing Lab Department of Computer Science and Technology School of InformaticsXiamen University Media Analytics and Computing Lab Department of Artificial Intelligence School of InformaticsXiamen University School of Automation Science and Electrical Engineering Beihang University Tencent Youtu Lab School of Software Tsinghua University

Graph convolutional networks(GCNs) provide a promising way to extract the useful information from graph-structured data. Most of the existing GCNs methods usually focus on local neighborhood information based on specific convolution operations, and ignore the global structure of the input data. To extract the latent representation for the graph-structured data more effectively, we introduce a deepwalk strategy into GCNs to efficiently explore the global graph information. This strategy can complement the local neighborhood information of a graph, resulting in the more robust representation for the graph *** fusion of the local neighboring and global structured information of a graph can further facilitate deep feature learning at the output layer of GCNs for node classification. Experimental results show that the proposed model has achieved state-of-the-art results on three benchmark datasets including Cora, Citeseer,and Pubmed citation networks.

关键词： graph convolutional networks global information fusion node classification

来源：评论

学校读者我要写书评

暂无评论

Mirai: A Wearable Proactive AI "Inner-Voice" for Contextual Nudging

Mirai: A Wearable Proactive AI "Inner-Voice" for Contextual ...

引用

2025 CHI Conference on Human Factors in computing Systems, CHI EA 2025

作者： Fang, Cathy Mengying Samaradivakara, Yasith Maes, Pattie Nanayakkara, Suranga MIT Media Lab. Cambridge MA United States Augmented Human Lab. Department of Information Systems and Analytics National University of Singapore Singapore Singapore Augmented Human Lab. School of Computing National University of Singapore Singapore Singapore

ISBN: (纸本)9798400713958

People often find it difficult to turn their intentions into real actions—a challenge that affects both personal growth and mental well-being. While established methods like cognitive-behavioral therapy and mindfulness training help people become more aware of their behaviors and set clear goals, these approaches cannot provide immediate guidance when people fall into automatic reactions or habits. We introduce Mirai, a novel wearable AI system with an integrated camera, real-time speech processing, and personalized voice-cloning to provide proactive and contextual nudges for positive behavior change. Mirai continuously monitors and analyzes the user’s environment to anticipate their intentions, generating contextually-appropriate responses delivered in the user’s own cloned voice. We demonstrate the application of Mirai through three scenarios focusing on dietary choices, work productivity, and communication skills. We also discuss future work on improving the proactive agent via human feedback and the need for a longitudinal study in naturalistic settings. © 2025 Copyright held by the owner/author(s).

关键词： context-aware generative ai goals nudging proactive agents voice

来源：评论

学校读者我要写书评

暂无评论

Panel Submission: Cyberinfrastructure Innovations Supporting Research, Equity, Discovery 24

Panel Submission: Cyberinfrastructure Innovations Supporting...

引用

25th Annual International Conference on Digital Government Research, DGO 2024

作者： Knepper, Rich Kee, Kerk F. Chen, Yu-Che Xu, Shan Cornell Center for Advanced Computing Cornell University United States College of Media and Communication Texas Tech University Texas United States School of Public Administration/Digital Governance and Analytics Lab College of Public Affairs and Community Service University of Nebraska at Omaha OmahaNE United States

ISBN: (纸本)9798400709883

Cyberinfrastructure refers to the computational, data, networking and human resources that support scientific research. Frequently cyberinfrastructure is a strategic, nation-level investment in research and development resources that has long-term impacts in national competitiveness, innovation, and discovery. As such, cyberinfrastructure is an essential element of a cohesive national science policy. Our panel discusses the development and implementation of cyberinfrastructure resources that support science;how their management and makeup affects performance;metrics for assessment;and how well cyberinfrastructure service delivery reaches equity goals. © 2024 Copyright held by the owner/author(s).

关键词： Cyberinfrastructure digital innovation platform policy informatics science policy

来源：评论

学校读者我要写书评

暂无评论

Self-Paced Partial Domain-Aware Learning for Face Anti-Spoofing

Self-Paced Partial Domain-Aware Learning for Face Anti-Spoof...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Zhiyi Chen Yao Lu Xinzhe Deng Jia Meng Shengchuan Zhang Liujuan Cao Department of Artificial Intelligence School of Informatics Media Analytics and Computing Lab Xiamen University China Youtu Lab Tencent China

With the widespread deployment of face authentication systems, domain generalization (DG) based face anti-spoofing (FAS) security approaches have drawn growing attention. Existing generalization-based methods always attempt to extract domain-invariant task information from data and eliminate domain-dependent information from representation space. However, they neglect that domain-related information may also contain helpful features for the classification task. To address this issue, we propose a self-paced partial domain-aware framework (SPDA) to preserve domain-related features helpful for the discrimination of fake and real faces, thereby increasing generalization for unseen domains. Specifically, a training strategy based on contrastive learning is adopted to construct domain-adapted and domain-aware task-related representation spaces. Then, a partial domain-aware adaptation module (PDA) is proposed to preserve valuable domain-related information for the task features that the network considers useful for mixture-domain classification. In addition, the proposed self-paced method(SCM) continuously explores potential clusters with insufficient representation to enhance further the feature extractor’s capability and the effectiveness of the PDA module. Extensive experiments demonstrate the effectiveness of our method compared to SOTA algorithms.

关键词： Training Signal processing algorithms Clustering algorithms Signal processing Feature extraction Security Data mining

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：