圖上的對(duì)抗與攻擊精選論文列表(?2021相關(guān)論文一覽)
來(lái)源：深度學(xué)習(xí)與圖網(wǎng)絡(luò) 本文約1400字，建議閱讀5分鐘
本文為你分享圖上的對(duì)抗與攻擊精選論文。
大規(guī)模攻擊圖神經(jīng)網(wǎng)絡(luò) 圖神經(jīng)網(wǎng)絡(luò)的黑盒梯度攻擊： 更深入洞察圖的攻擊和防御 增強(qiáng)多路復(fù)用網(wǎng)絡(luò)對(duì)節(jié)點(diǎn)社區(qū)級(jí)聯(lián)故障的魯棒性和彈性 PATHATTACK： 攻擊復(fù)雜網(wǎng)絡(luò)中的最短路徑 Deformable shape的通用譜對(duì)抗攻擊 Preserve, Promote, or Attack?通過(guò)拓?fù)鋽_動(dòng)的 GNN 解釋 網(wǎng)絡(luò)嵌入攻擊： 一種基于歐幾里德距離的方法 通過(guò)監(jiān)督網(wǎng)絡(luò)Poisoning對(duì)網(wǎng)絡(luò)嵌入的對(duì)抗性攻擊 DeHiB： 通過(guò)對(duì)抗性擾動(dòng)對(duì)半監(jiān)督學(xué)習(xí)的深層隱藏后門(mén)攻擊 GraphAttacker： 一個(gè)通用的多任務(wù)圖攻擊框架 圖神經(jīng)網(wǎng)絡(luò)的成員推理攻擊
Attacking Graph Neural Networks at Scale
Black-box Gradient Attack on Graph Neural Networks: Deeper Insights in Graph-based Attack and Defense
Enhancing Robustness and Resilience of Multiplex Networks Against Node-Community Cascading Failures
PATHATTACK: Attacking Shortest Paths in Complex Networks
Universal Spectral Adversarial Attacks for Deformable Shapes
Preserve, Promote, or Attack? GNN Explanation via Topology Perturbation
Network Embedding Attack: An Euclidean Distance Based Method
Adversarial Attack on Network Embeddings via Supervised Network Poisoning
DeHiB: Deep Hidden Backdoor Attack on Semi-Supervised Learning via Adversarial Perturbation
GraphAttacker: A General Multi-Task Graph Attack Framework
Membership Inference Attack on Graph Neural Networks
圖神經(jīng)網(wǎng)絡(luò)的對(duì)抗性標(biāo)簽翻轉(zhuǎn)攻擊和防御 對(duì)圖神經(jīng)網(wǎng)絡(luò)的探索性對(duì)抗攻擊 對(duì)圖卷積網(wǎng)絡(luò)的有針對(duì)性的通用攻擊 在不改變現(xiàn)有連接的情況下攻擊基于圖的分類(lèi) 學(xué)習(xí)通過(guò)有針對(duì)性的擾動(dòng)欺騙知識(shí)圖譜增強(qiáng)模型 基于圖神經(jīng)網(wǎng)絡(luò)的時(shí)空預(yù)測(cè)的一種頂點(diǎn)攻擊 欺騙圖神經(jīng)網(wǎng)絡(luò)的單節(jié)點(diǎn)攻擊 圖神經(jīng)網(wǎng)絡(luò)的黑盒對(duì)抗攻擊作為影響最大化問(wèn)題 深度圖匹配的對(duì)抗性攻擊 對(duì)圖神經(jīng)網(wǎng)絡(luò)進(jìn)行Practical對(duì)抗性攻擊 一種對(duì)隱私保護(hù)記錄鏈接的圖匹配攻擊 通過(guò) GAN 對(duì)圖嵌入的自適應(yīng)對(duì)抗性攻擊 乘法器交替方向法對(duì)圖神經(jīng)網(wǎng)絡(luò)的可擴(kuò)展對(duì)抗性攻擊 針對(duì)用于惡意軟件檢測(cè)的圖神經(jīng)網(wǎng)絡(luò)的語(yǔ)義保留強(qiáng)化學(xué)習(xí)攻擊 對(duì)大規(guī)模圖的對(duì)抗性攻擊 通過(guò)影響函數(shù)(Influence Function)對(duì)圖神經(jīng)網(wǎng)絡(luò)進(jìn)行有效的規(guī)避攻擊 基于強(qiáng)化學(xué)習(xí)的黑盒規(guī)避攻擊在動(dòng)態(tài)圖中進(jìn)行鏈接預(yù)測(cè) 針對(duì)無(wú)標(biāo)度網(wǎng)絡(luò)的 BC 分類(lèi)的對(duì)抗性攻擊 基于圖神經(jīng)網(wǎng)絡(luò)的鏈路預(yù)測(cè)算法的對(duì)抗性攻擊 圖神經(jīng)網(wǎng)絡(luò)的Practical對(duì)抗性攻擊 通過(guò)迭代梯度攻擊的鏈路預(yù)測(cè)對(duì)抗性攻擊 對(duì)圖結(jié)構(gòu)化數(shù)據(jù)的有效對(duì)抗性攻擊 圖Backdoor 圖神經(jīng)網(wǎng)絡(luò)的Backdoor攻擊 通過(guò) Nash 強(qiáng)化學(xué)習(xí)進(jìn)行垃圾郵件發(fā)送檢測(cè) 圖神經(jīng)網(wǎng)絡(luò)的對(duì)抗性攻擊：擾動(dòng)及其模式 對(duì)分層圖池化神經(jīng)網(wǎng)絡(luò)的對(duì)抗性攻擊 從圖神經(jīng)網(wǎng)絡(luò)竊取鏈接 通過(guò)注入惡意節(jié)點(diǎn)對(duì)圖數(shù)據(jù)進(jìn)行可擴(kuò)展攻擊 網(wǎng)絡(luò)中斷：最大化社交網(wǎng)絡(luò)中的分歧和兩極分化 網(wǎng)絡(luò)中意見(jiàn)動(dòng)態(tài)的對(duì)抗性擾動(dòng) 圖神經(jīng)網(wǎng)絡(luò)上的非目標(biāo)特定節(jié)點(diǎn)注入攻擊：一種分層強(qiáng)化學(xué)習(xí)方法 MGA：網(wǎng)絡(luò)上的動(dòng)量梯度攻擊 通過(guò)對(duì)圖卷積網(wǎng)絡(luò)進(jìn)行Poisoning鄰居的間接對(duì)抗性攻擊 圖通用對(duì)抗性攻擊：一些不良行為者破壞圖學(xué)習(xí)模型 對(duì)無(wú)標(biāo)度網(wǎng)絡(luò)的對(duì)抗性攻擊：測(cè)試物理標(biāo)準(zhǔn)的穩(wěn)健性 通過(guò)隱藏個(gè)人對(duì)社區(qū)檢測(cè)的對(duì)抗性攻擊
Adversarial Label-Flipping Attack and Defense for Graph Neural Networks
Exploratory Adversarial Attacks on Graph Neural Networks
A Targeted Universal Attack on Graph Convolutional Network
Attacking Graph-Based Classification without Changing Existing Connections
Learning to Deceive Knowledge Graph Augmented Models via Targeted Perturbation
One Vertex Attack on Graph Neural Networks-based Spatiotemporal Forecasting
Single-Node Attack for Fooling Graph Neural Networks
Black-Box Adversarial Attacks on Graph Neural Networks as An Influence Maximization Problem
Adversarial Attacks on Deep Graph Matching | Attack | Graph Matching | Deep Graph Matching Models
Towards More Practical Adversarial Attacks on Graph Neural Networks
A Graph Matching Attack on Privacy-Preserving Record Linkage
Adaptive Adversarial Attack on Graph Embedding via GAN
Scalable Adversarial Attack on Graph Neural Networks with Alternating Direction Method of Multipliers
Semantic-preserving Reinforcement Learning Attack Against Graph Neural Networks for Malware Detection
Adversarial Attack on Large Scale Graph
Efficient Evasion Attacks to Graph Neural Networks via Influence Function
Reinforcement Learning-based Black-Box Evasion Attacks to Link Prediction in Dynamic Graphs
Adversarial attack on BC classification for scale-free networks
Adversarial Attacks on Link Prediction Algorithms Based on Graph Neural Networks
Practical Adversarial Attacks on Graph Neural Networks
Link Prediction Adversarial Attack Via Iterative Gradient Attack
An Efficient Adversarial Attack on Graph Structured Data
Graph Backdoor | Attack | Node Classification Graph Classification
Backdoor Attacks to Graph Neural Networks
Robust Spammer Detection by Nash Reinforcement Learning
Adversarial Attacks on Graph Neural Networks: Perturbations and their Patterns
Adversarial Attack on Hierarchical Graph Pooling Neural Networks
Stealing Links from Graph Neural Networks
Scalable Attack on Graph Data by Injecting Vicious Nodes
Network disruption: maximizing disagreement and polarization in social networks
Adversarial Perturbations of Opinion Dynamics in Networks
Non-target-specific Node Injection Attacks on Graph Neural Networks: A Hierarchical Reinforcement Learning Approach
MGA: Momentum Gradient Attack on Network | Attack | Node Classification, Community Detection
Indirect Adversarial Attacks via Poisoning Neighbors for Graph Convolutional Networks
Graph Universal Adversarial Attacks: A Few Bad Actors Ruin Graph Learning Models
Adversarial Attacks to Scale-Free Networks: Testing the Robustness of Physical Criteria
Adversarial Attack on Community Detection by Hiding Individuals
編輯：文婧