检索结果-内蒙古大学图书馆

IEEE computer Society Conference on computer Vision and Pattern Recognition Workshops (CVPRW)

作者： Xinwei Zhang Tianyuan Zhang Yitong Zhang Shuangcheng Liu School of Computer Science and Engineering Beihang University Beijing China State Key Lab of Software Development Environment Beihang University Beijing China Shen Yuan Honors College Beihang University Beijing China

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Deep neural networks are susceptible to attacks from adversarial examples in recent years. Especially, the black-box attacks cause a more serious threat to practical applications. However, while most existing black-box attacks have achieved a high success rate in deceiving models, they have not focused on the stealthiness of adversarial examples, often exhibiting suspicious visual appearances. To address this issue, this paper proposes the Mask Momentum Iterative Attack (MMIA), which introduces a masking mechanism and adopts an optimal perturbation strategy to identify regions of an image most vulnerable to attacks. This approach effectively ensures the transferability and stealthiness of adversarial examples. Simultaneously, by integrating image enhancement techniques and temporal and spatial momentum terms into the iterative process of the attack, we prevent the attack from getting stuck in local optima, further improving the transferability of adversarial examples. To enhance the success rate of black-box attacks, we apply MMIA to a model ensemble using a joint optimization strategy. We demonstrate that adversarially trained models with a strong defense ability are also susceptible to our black-box attacks. We conduct extensive experiments on classification tasks using common vision models, and our results significantly demonstrate the superiority of our method over state-of-the-art approaches when considering both transferability and stealthiness.

关键词： Visualization Fuses Computational modeling Perturbation methods Closed box Pattern recognition Iterative methods

来源：评论

学校读者我要写书评

暂无评论

Differentially Private Generative Model with Ratio-Based Gradient Clipping 4th

Differentially Private Generative Model with Ratio-Based Gr...

引用

4th International Conference on Machine Learning for Cyber Security, ML4CS 2022

作者： Lin, Jianchen Yao, Yanqing State Key Laboratory of Software Development Environment Beihang University Beijing100191 China State Key Laboratory of Cryptology Beijing100878 China Key Laboratory of Aerospace Network Security Ministry of Industry and Information Technology School of Cyber Science and Technology Beihang University Beijing100191 China State Grid Fujian Fuzhou350100 China

ISBN: (纸本)9783031200953

As a branch of deep learning, conditional Generative Adversarial Net (cGAN) has achieved remarkable results in target-oriented image generation, but the process of training the model lacks privacy guarantee. Differential Privacy (DP) is a privacy-preserving mechanism that could be utilized in the Stochastic Gradient Descent (SGD) of deep learning to maintain the confidentiality of training data. However, the quality and diversity of differentially private conditional image synthesis remain large room for improvement because traditional mechanisms with thick granularities and rigid clipping bounds in Differentially Private SGD (DPSGD) could lead to huge performance loss. To tackle this, we propose three algorithms with different clipping granularities and clipping bound distributions, and investigate the effects on enhancing conditional image generation under Differentially Private Projection cGAN (DP-PcGAN). We compare and analyze the performance of our strategies with methods concerned in terms of seven metrics, finding ratio-based gradient clipping highly effective and robust in training DP-PcGAN for generating visually promising images. Specifically, the Per-layer RGC surpasses previous methods under all evaluation metrics utilizing the second least privacy expense for MNIST dataset, while Per-unit RGC attains encouraging outcomes with a moderate converging speed under FashionMNIST dataset. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Further Explanations on "SAT Requires Exhaustive Search"

arXiv

引用

arXiv 2024年

作者： Dong, Qingxiu Zhou, Guangyan Xu, Ke School of Computer Science Peking University China Department of Mathematics and Statistics Beijing Technology and Business University China State Key Lab of Software Development Environment Beihang University China

Recently, Xu and Zhou [2023] introduced a constructive approach for exploring computational hardness, proving that SAT requires exhaustive search. In light of certain misinterpretations concerning the contributions and proofs in that paper, we focus on providing detailed explanations in this work. We begin by delineating the core innovation of the constructive approach, shedding light on the pivotal concept of algorithm designability. We address the overlooked white-box diagonalization method and highlight the concept of an almost independent solution space. In response to specific misunderstandings, such as the concerns surrounding the assumptions of Lemma 3.1, we offer comprehensive clarifications aimed at improving the comprehension of the proof. We are grateful for the feedback received on our prior paper and hope this work can foster a more well-informed discussion. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

ATA-Cache: Contention Mitigation for GPU Shared L1 Cache with Aggregated Tag Array

arXiv

引用

arXiv 2023年

作者： Xu, Xiangrong Wang, Liang Xiao, Limin Liu, Lei Xie, Xilong Han, Meng Liu, Hao State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beihang University Beijing100191 China

GPU shared L1 cache is a promising architecture while still suffering from high resource contentions. We present a GPU shared L1 cache architecture with an aggregated tag array that minimizes the L1 cache contentions and takes full advantage of inter-core locality. The key idea is to decouple and aggregate the tag arrays of multiple L1 caches so that the cache requests can be compared with all tag arrays in parallel to probe the replicated data in other caches. The GPU caches are only accessed by other GPU cores when replicated data exists, filtering out unnecessary cache accesses that cause high resource contentions. The experimental results show that GPU IPC can be improved by 12% on average for applications with a high inter-core locality. Copyright © 2023, The Authors. All rights reserved.

关键词： Graphics processing unit

来源：评论

学校读者我要写书评

暂无评论

Personalized Federated Learning System Based on Permissioned Blockchain

Personalized Federated Learning System Based on Permissioned...

引用

2021 International Conference on Intelligent Computing, Automation and Systems, ICICAS 2021

作者： Yuan, Bo Qiu, Wangjie Beihang University School of Computer Science and Engineering State Key Laboratory of Software Development Environment Beijing100191 China Beihang University State Key Laboratory of Software Development Environment Beijing100191 China

ISBN: (纸本)9781665428101

Federated learning ensures the privacy of data generated by large-scale IoT devices. Existing federated learning frameworks, based on centralized model coordinators, still face serious security challenges such as single point of failure and lack of privacy. In this paper, we propose a personalized federated learning system based on permissioned blockchain, which is divided into four layers of architecture are IoT device layer, network layer, edge computing layer, blockchain layer, and application layer, using permissioned blockchain as a federated learning server. And a permission blockchain-based personalized federation learning algorithm is proposed, which can achieve privacy protection and resistance to poisoning attacks with high accuracy. The experimental results revealed that the system has high privacy protection and anti-poisoning attack capability and can be deployed in edge computing situations. © 2021 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

FuseFPS: Accelerating Farthest Point Sampling with Fusing KD-tree Construction for Point Clouds

FuseFPS: Accelerating Farthest Point Sampling with Fusing KD...

引用

Asia and South Pacific Design Automation Conference

作者： Meng Han Liang Wang Limin Xiao Hao Zhang Chenhao Zhang Xilong Xie Shuai Zheng Jin Dong State Key Laboratory of Software Development Environment Beijing China School of Computer Science and Engineering Beihang University Beijing China Beijing Academy of Blockchain and Edge Computing Beijing China

Point cloud analytics has become a critical workload for embedded and mobile platforms across various applications. Farthest point sampling (FPS) is a fundamental and widely used kernel in point cloud processing. However, the heavy external memory access makes FPS a performance bottleneck for real-time point cloud processing. Although bucket-based farthest point sampling can significantly reduce unnecessary memory accesses during the point sampling stage, the KD-tree construction stage becomes the predominant contributor to execution time. In this paper, we present FuseFPS, an architecture and algorithm co-design for bucket-based farthest point sampling. We first propose a hardware-friendly sampling-driven KD-tree construction algorithm. The algorithm fuses the KD-tree construction stage into the point sampling stage, further reducing memory accesses. Then, we design an efficient accelerator for bucket-based point sampling. The accelerator can offload the entire bucket-based FPS kernel at a low hardware cost. Finally, we evaluate our approach on various point cloud datasets. The detailed experiments show that compared to the state-of-the-art accelerator QuickFPS, FuseFPS achieves about $4.3\times $ and about $6.1\times $ improvements on speed and power efficiency, respectively.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Parameter Optimization via Reinforcement Learning for the Regulation of Swarms 9

Parameter Optimization via Reinforcement Learning for the Re...

引用

9th International Conference on Information, Cybernetics, and Computational Social Systems, ICCSS 2023

作者： Wu, Qizhen Liu, Gaoxiang Liu, Kexin Chen, Lei School of Automation Science and Electrical Engineering Beihang University Beijing China Environment School of Automation Science and Electrical Engineering Beihang University The State Key Laboratory of Software Development Beijing China Advanced Research Institute of Multidisciplinary Sciences Beijing Institute of Technology Beijing China Yangtze Delta Region Academy of Beijing Institute of Technology Jiaxing China

ISBN: (纸本)9798350342239

The bird-oid object (Boids) model proposes a control algorithm to make the positions between agents achieve cooperative stability. By changing the parameters of cohesion and repulsion in the algorithm, the agents in the swarm can be made to converge to different positions, causing expansion and contraction of the formation. But it is often more difficult to select the appropriate parameters to form the ideal formation. Therefore, this paper proposes a method to improve the cohesive and repulsive parameters in the Boids model based on Q-learning network to achieve a simulation scenario with continuous obstacle avoidance and maximum coverage of space. © 2023 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Sim-to-Real Grasp Detection with Global-to-Local RGB-D Adaptation

Sim-to-Real Grasp Detection with Global-to-Local RGB-D Adapt...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Haoxiang Ma Ran Qin Modi Shi Boyang Gao Di Huang State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beihang University Beijing China Geometry Robotics and the School of Computer Science and Technology Harbin Institute of Technology Harbin China

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

This paper focuses on the sim-to-real issue of RGB-D grasp detection and formulates it as a domain adaptation problem. In this case, we present a global-to-local method to address hybrid domain gaps in RGB and depth data and insufficient multi-modal feature alignment. First, a self-supervised rotation pre-training strategy is adopted to deliver robust initialization for RGB and depth networks. We then propose a global-to-local alignment pipeline with individual global domain classifiers for scene features of RGB and depth images as well as a local one specifically working for grasp features in the two modalities. In particular, we propose a grasp prototype adaptation module, which aims to facilitate fine-grained local feature alignment by dynamically updating and matching the grasp prototypes from the simulation and real-world scenarios throughout the training process. Due to such designs, the proposed method substantially reduces the domain shift and thus leads to consistent performance improvements. Extensive experiments are conducted on the GraspNet-Planar benchmark and physical environment, and superior results are achieved which demonstrate the effectiveness of our method. Code is available at https://***/mahaoxiang822/GL-MSDA.

关键词： Training Codes Pipelines Prototypes Benchmark testing Robustness Robotics and automation

来源：评论

学校读者我要写书评

暂无评论

Sim-to-Real Grasp Detection with Global-to-Local RGB-D Adaptation

arXiv

引用

arXiv 2024年

作者： Ma, Haoxiang Qin, Ran Shi, Modi Gao, Boyang Huang, Di The State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beihang University Beijing China The Geometry Robotics and the School of Computer Science and Technology Harbin Institute of Technology Harbin China

关键词：

来源：评论

学校读者我要写书评

暂无评论

RoNID: New Intent Discovery with Generated-Reliable labels and Cluster-friendly Representations

arXiv

引用

arXiv 2024年

作者： Zhang, Shun Yan, Chaoran Yang, Jian Ren, Changyu Bai, Jiaqi Li, Tongliang Li, Zhoujun State Key Lab of Software Development Environment Beihang University China School of Cyber Science and Technology Beihang University China Beijing Information Science and Technology University China

New Intent Discovery (NID) strives to identify known and reasonably deduce novel intent groups in the open-world scenario. But current methods face issues with inaccurate pseudo-labels and poor representation learning, creating a negative feedback loop that degrades overall model performance, including accuracy and the adjusted rand index. To address the aforementioned challenges, we propose a Robust New Intent Discovery (RoNID) framework optimized by an EM-style method, which focuses on constructing reliable pseudo-labels and obtaining cluster-friendly discriminative representations. RoNID comprises two main modules: reliable pseudo-label generation module and cluster-friendly representation learning module. Specifically, the pseudo-label generation module assigns reliable synthetic labels by solving an optimal transport problem in the E-step, which effectively provides high-quality supervised signals for the input of the cluster-friendly representation learning module. To learn cluster-friendly representation with strong intra-cluster compactness and large inter-cluster separation, the representation learning module combines intra-cluster and inter-cluster contrastive learning in the M-step to feed more discriminative features into the generation module. RoNID can be performed iteratively to ultimately yield a robust model with reliable pseudo-labels and cluster-friendly representations. Experimental results on multiple benchmarks demonstrate our method brings substantial improvements over previous state-of-the-art methods by a large margin of +1∼+4 points. Copyright © 2024, The Authors. All rights reserved.

关键词： Iterative methods

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：