FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313229355> ?p ?o ?g. }

• W4313229355 endingPage "892" @default.
• W4313229355 startingPage "877" @default.
• W4313229355 abstract "Deep neural network-based learning methods have been considered promising techniques used in beam selection problems. However, existing research ignores the peculiar vulnerabilities of neural networks. The adversaries can use data poisoning to embed predefined triggers into a model during training time such that the neural network-based beam model may make an incorrect output decision of a test example when patched with the trigger. Data poisoning offers attackers the possibility to build backdoors. The goal of backdoors is often unethical, such as giving users a poor experience by manipulating infected models to output inappropriate beams. In this paper, first, we introduce a simple backdoor attack method by using data poisoning in a mmWave beam selection system. By numerical simulations, we verify that this poisoning attack is effective for neural networks with different structures. In addition, we explore the effect of poisoned data volume on the effect of backdoor attacks. The results show that the backdoor can be successfully implanted into the beam selection neural network. Besides, we fine-tune the trained model for a new wireless communication environment, and the results show that backdoors still exist even when the model is tuned with data from new scenarios. Then, we propose a machine unlearning solution to mitigate the backdoor of the trained beam selection model. The problem of eliminating backdoors is modeled as a minimax optimization problem. We propose a novel adversarial unlearning method along with label smoothing to solve the backdoor removal problem. We compared the proposed backdoor elimination method with the classical fine-tuning elimination method and the neural network pruning method through numerical simulations. The results show that the fine-tuning and the pruning methods cannot effectively remove the backdoor. The proposed machine unlearning method can make the trained model forget about the backdoor under the condition that the performance of the benign task (beam selection tasks when the trigger does not appear) is guaranteed to be slightly degraded. In summary, our work illustrates that data poisoning-based backdoor attacks may exist in wireless networks, and we propose a scheme to eliminate backdoors." @default.
• W4313229355 created "2023-01-06" @default.
• W4313229355 creator A5019360899 @default.
• W4313229355 creator A5045457613 @default.
• W4313229355 creator A5056225611 @default.
• W4313229355 creator A5056820584 @default.
• W4313229355 creator A5083864182 @default.
• W4313229355 date "2023-02-01" @default.
• W4313229355 modified "2023-10-13" @default.
• W4313229355 title "Poison Neural Network-Based mmWave Beam Selection and Detoxification With Machine Unlearning" @default.
• W4313229355 cites W2041287972 @default.
• W4313229355 cites W2166486216 @default.
• W4313229355 cites W2183341477 @default.
• W4313229355 cites W2614851539 @default.
• W4313229355 cites W2792845087 @default.
• W4313229355 cites W2807363941 @default.
• W4313229355 cites W2898434483 @default.
• W4313229355 cites W2907201671 @default.
• W4313229355 cites W2914111477 @default.
• W4313229355 cites W2962944637 @default.
• W4313229355 cites W2963542245 @default.
• W4313229355 cites W2969424089 @default.
• W4313229355 cites W2970203070 @default.
• W4313229355 cites W2983729222 @default.
• W4313229355 cites W2995099986 @default.
• W4313229355 cites W2995240473 @default.
• W4313229355 cites W2998881388 @default.
• W4313229355 cites W3031954909 @default.
• W4313229355 cites W3105281589 @default.
• W4313229355 cites W3108957811 @default.
• W4313229355 cites W3119458351 @default.
• W4313229355 cites W3130401643 @default.
• W4313229355 cites W3135588948 @default.
• W4313229355 cites W3154155772 @default.
• W4313229355 cites W3154878987 @default.
• W4313229355 cites W3184974140 @default.
• W4313229355 cites W3186724161 @default.
• W4313229355 cites W3193643066 @default.
• W4313229355 cites W3199812977 @default.
• W4313229355 cites W3206552804 @default.
• W4313229355 cites W3208304128 @default.
• W4313229355 cites W3212590122 @default.
• W4313229355 cites W3213965409 @default.
• W4313229355 cites W4225688943 @default.
• W4313229355 cites W4225750584 @default.
• W4313229355 cites W4230309686 @default.
• W4313229355 cites W4286256876 @default.
• W4313229355 doi "https://doi.org/10.1109/tcomm.2022.3232794" @default.
• W4313229355 hasPublicationYear "2023" @default.
• W4313229355 type Work @default.
• W4313229355 citedByCount "0" @default.
• W4313229355 crossrefType "journal-article" @default.
• W4313229355 hasAuthorship W4313229355A5019360899 @default.
• W4313229355 hasAuthorship W4313229355A5045457613 @default.
• W4313229355 hasAuthorship W4313229355A5056225611 @default.
• W4313229355 hasAuthorship W4313229355A5056820584 @default.
• W4313229355 hasAuthorship W4313229355A5083864182 @default.
• W4313229355 hasConcept C119857082 @default.
• W4313229355 hasConcept C154945302 @default.
• W4313229355 hasConcept C2781045450 @default.
• W4313229355 hasConcept C38652104 @default.
• W4313229355 hasConcept C41008148 @default.
• W4313229355 hasConcept C50644808 @default.
• W4313229355 hasConcept C81917197 @default.
• W4313229355 hasConceptScore W4313229355C119857082 @default.
• W4313229355 hasConceptScore W4313229355C154945302 @default.
• W4313229355 hasConceptScore W4313229355C2781045450 @default.
• W4313229355 hasConceptScore W4313229355C38652104 @default.
• W4313229355 hasConceptScore W4313229355C41008148 @default.
• W4313229355 hasConceptScore W4313229355C50644808 @default.
• W4313229355 hasConceptScore W4313229355C81917197 @default.
• W4313229355 hasFunder F4320321001 @default.
• W4313229355 hasFunder F4320335777 @default.
• W4313229355 hasFunder F4320335787 @default.
• W4313229355 hasIssue "2" @default.
• W4313229355 hasLocation W43132293551 @default.
• W4313229355 hasOpenAccess W4313229355 @default.
• W4313229355 hasPrimaryLocation W43132293551 @default.
• W4313229355 hasRelatedWork W2386387936 @default.
• W4313229355 hasRelatedWork W2961085424 @default.
• W4313229355 hasRelatedWork W3046775127 @default.
• W4313229355 hasRelatedWork W4205958290 @default.
• W4313229355 hasRelatedWork W4285260836 @default.
• W4313229355 hasRelatedWork W4286629047 @default.
• W4313229355 hasRelatedWork W4306321456 @default.
• W4313229355 hasRelatedWork W4306674287 @default.
• W4313229355 hasRelatedWork W1629725936 @default.
• W4313229355 hasRelatedWork W4224009465 @default.
• W4313229355 hasVolume "71" @default.
• W4313229355 isParatext "false" @default.
• W4313229355 isRetracted "false" @default.
• W4313229355 workType "article" @default.