000 05701nam a22005655i 4500
001 978-981-97-0425-5
003 DE-He213
005 20240730172355.0
007 cr nn 008mamaa
008 240603s2024 si | s |||| 0|eng d
020 _a9789819704255
_9978-981-97-0425-5
024 7 _a10.1007/978-981-97-0425-5
_2doi
050 4 _aQ334-342
050 4 _aTA347.A78
072 7 _aUYQ
_2bicssc
072 7 _aCOM004000
_2bisacsh
072 7 _aUYQ
_2thema
082 0 4 _a006.3
_223
100 1 _aChen, Jinyin.
_eauthor.
_4aut
_4http://id.loc.gov/vocabulary/relators/aut
_9103326
245 1 0 _aAttacks, Defenses and Testing for Deep Learning
_h[electronic resource] /
_cby Jinyin Chen, Ximin Zhang, Haibin Zheng.
250 _a1st ed. 2024.
264 1 _aSingapore :
_bSpringer Nature Singapore :
_bImprint: Springer,
_c2024.
300 _aXX, 399 p. 128 illus., 126 illus. in color.
_bonline resource.
336 _atext
_btxt
_2rdacontent
337 _acomputer
_bc
_2rdamedia
338 _aonline resource
_bcr
_2rdacarrier
347 _atext file
_bPDF
_2rda
505 0 _aPerturbation Optimized Black-Box Adversarial Attacks via Genetic Algorithm -- Feature Transfer Based Stealthy Poisoning Attack for DNNs -- Adversarial Attacks on GNN Based Vertical Federated Learning -- A Novel DNN Object Contour Attack on Image Recognition -- Query-Efficient Adversarial Attack Against Vertical Federated Graph Learning -- Targeted Label Adversarial Attack on Graph Embedding -- Backdoor Attack on Dynamic Link Prediction -- Attention Mechanism based Adversarial Attack against DRL -- Characterizing Adversarial Examples via Local Gradient Checking -- A Novel Adversarial Defense by Refocusing on Critical Areas -- Neuron-level Inverse Perturbation Against Adversarial Attacks -- Adaptive Channel Transformation-based Detector for Adversarial Attacks -- Defense Against Free-rider Attack From the Weight Evolving Frequency -- An Effective Model Copyright Protection for Federated Learning -- Guard the vertical federated graph learning from Property Inference Attack -- Using Adversarial Examples to Against Backdoor Attack in FL -- Evaluating the Adversarial Robustness of Deep Model by Decision Boundaries -- Certifiable Prioritization for Deep Neural Networks via Movement Cost in Feature Space -- Interpretable White-Box Fairness Testing through Biased Neuron Identification -- A Deep Learning Framework for Dynamic Network Link Prediction. .
520 _aThis book provides a systematic study on the security of deep learning. With its powerful learning ability, deep learning is widely used in CV, FL, GNN, RL, and other scenarios. However, during the process of application, researchers have revealed that deep learning is vulnerable to malicious attacks, which will lead to unpredictable consequences. Take autonomous driving as an example, there were more than 12 serious autonomous driving accidents in the world in 2018, including Uber, Tesla and other high technological enterprises. Drawing on the reviewed literature, we need to discover vulnerabilities in deep learning through attacks, reinforce its defense, and test model performance to ensure its robustness. Attacks can be divided into adversarial attacks and poisoning attacks. Adversarial attacks occur during the model testing phase, where the attacker obtains adversarial examples by adding small perturbations. Poisoning attacks occur during the model training phase, where the attacker injects poisoned examples into the training dataset, embedding a backdoor trigger in the trained deep learning model. An effective defense method is an important guarantee for the application of deep learning. The existing defense methods are divided into three types, including the data modification defense method, model modification defense method, and network add-on method. The data modification defense method performs adversarial defense by fine-tuning the input data. The model modification defense method adjusts the model framework to achieve the effect of defending against attacks. The network add-on method prevents the adversarial examples by training the adversarial example detector. Testing deep neural networks is an effective method to measure the security and robustness of deep learning models. Through test evaluation, security vulnerabilities and weaknesses in deep neural networks can be identified. By identifying and fixing these vulnerabilities, the security and robustness of the model can be improved. Our audience includes researchers in the field of deep learning security, as well as software development engineers specializing in deep learning.
650 0 _aArtificial intelligence.
_93407
650 0 _aComputer engineering.
_910164
650 0 _aComputer networks .
_931572
650 0 _aNeural networks (Computer science) .
_9103328
650 1 4 _aArtificial Intelligence.
_93407
650 2 4 _aComputer Engineering and Networks.
_9103329
650 2 4 _aMathematical Models of Cognitive Processes and Neural Networks.
_932913
700 1 _aZhang, Ximin.
_eauthor.
_4aut
_4http://id.loc.gov/vocabulary/relators/aut
_9103331
700 1 _aZheng, Haibin.
_eauthor.
_4aut
_4http://id.loc.gov/vocabulary/relators/aut
_9103332
710 2 _aSpringerLink (Online service)
_9103335
773 0 _tSpringer Nature eBook
776 0 8 _iPrinted edition:
_z9789819704248
776 0 8 _iPrinted edition:
_z9789819704262
776 0 8 _iPrinted edition:
_z9789819704279
856 4 0 _uhttps://doi.org/10.1007/978-981-97-0425-5
912 _aZDB-2-SCS
912 _aZDB-2-SXCS
942 _cEBK
999 _c88250
_d88250