Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386072370> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W4386072370 abstract "Since training a deep neural network (DNN) is costly, the well-trained deep models can be regarded as valuable intellectual property (IP) assets. The IP protection associated with deep models has been receiving increasing attentions in recent years. Passport-based method, which replaces normalization layers with passport layers, has been one of the few protection solutions that are claimed to be secure against advanced attacks. In this work, we tackle the issue of evaluating the security of passport-based IP protection methods. We propose a novel and effective ambiguity attack against passport-based method, capable of successfully forging multiple valid passports with a small training dataset. This is accomplished by inserting a specially designed accessory block ahead of the passport parameters. Using less than 10% of training data, with the forged passport, the model exhibits almost indistinguishable performance difference (less than 2%) compared with that of the authorized passport. In addition, it is shown that our attack strategy can be readily generalized to attack other IP protection methods based on watermark embedding. Directions for potential remedy solutions are also given." @default.
- W4386072370 created "2023-08-23" @default.
- W4386072370 creator A5010705747 @default.
- W4386072370 creator A5014934585 @default.
- W4386072370 creator A5036536809 @default.
- W4386072370 creator A5073693142 @default.
- W4386072370 date "2023-06-01" @default.
- W4386072370 modified "2023-09-27" @default.
- W4386072370 title "Effective Ambiguity Attack Against Passport-based DNN Intellectual Property Protection Schemes through Fully Connected Layer Substitution" @default.
- W4386072370 cites W2162517204 @default.
- W4386072370 cites W2194775991 @default.
- W4386072370 cites W2294710185 @default.
- W4386072370 cites W2568258226 @default.
- W4386072370 cites W2579318729 @default.
- W4386072370 cites W2768064608 @default.
- W4386072370 cites W2806082141 @default.
- W4386072370 cites W2935349488 @default.
- W4386072370 cites W2952608669 @default.
- W4386072370 cites W2964128659 @default.
- W4386072370 cites W2997717738 @default.
- W4386072370 cites W3043547428 @default.
- W4386072370 cites W3105676597 @default.
- W4386072370 cites W3156793535 @default.
- W4386072370 cites W3158240034 @default.
- W4386072370 cites W3159280306 @default.
- W4386072370 cites W3159937784 @default.
- W4386072370 cites W3168768313 @default.
- W4386072370 cites W3184974140 @default.
- W4386072370 cites W3205572447 @default.
- W4386072370 cites W3206880386 @default.
- W4386072370 cites W4206589716 @default.
- W4386072370 doi "https://doi.org/10.1109/cvpr52729.2023.00785" @default.
- W4386072370 hasPublicationYear "2023" @default.
- W4386072370 type Work @default.
- W4386072370 citedByCount "0" @default.
- W4386072370 crossrefType "proceedings-article" @default.
- W4386072370 hasAuthorship W4386072370A5010705747 @default.
- W4386072370 hasAuthorship W4386072370A5014934585 @default.
- W4386072370 hasAuthorship W4386072370A5036536809 @default.
- W4386072370 hasAuthorship W4386072370A5073693142 @default.
- W4386072370 hasConcept C108583219 @default.
- W4386072370 hasConcept C111472728 @default.
- W4386072370 hasConcept C111919701 @default.
- W4386072370 hasConcept C138885662 @default.
- W4386072370 hasConcept C154945302 @default.
- W4386072370 hasConcept C164112704 @default.
- W4386072370 hasConcept C189950617 @default.
- W4386072370 hasConcept C199360897 @default.
- W4386072370 hasConcept C2524010 @default.
- W4386072370 hasConcept C2777210771 @default.
- W4386072370 hasConcept C2780522230 @default.
- W4386072370 hasConcept C2984842247 @default.
- W4386072370 hasConcept C33923547 @default.
- W4386072370 hasConcept C34974158 @default.
- W4386072370 hasConcept C38652104 @default.
- W4386072370 hasConcept C41008148 @default.
- W4386072370 hasConcept C41608201 @default.
- W4386072370 hasConcept C50644808 @default.
- W4386072370 hasConceptScore W4386072370C108583219 @default.
- W4386072370 hasConceptScore W4386072370C111472728 @default.
- W4386072370 hasConceptScore W4386072370C111919701 @default.
- W4386072370 hasConceptScore W4386072370C138885662 @default.
- W4386072370 hasConceptScore W4386072370C154945302 @default.
- W4386072370 hasConceptScore W4386072370C164112704 @default.
- W4386072370 hasConceptScore W4386072370C189950617 @default.
- W4386072370 hasConceptScore W4386072370C199360897 @default.
- W4386072370 hasConceptScore W4386072370C2524010 @default.
- W4386072370 hasConceptScore W4386072370C2777210771 @default.
- W4386072370 hasConceptScore W4386072370C2780522230 @default.
- W4386072370 hasConceptScore W4386072370C2984842247 @default.
- W4386072370 hasConceptScore W4386072370C33923547 @default.
- W4386072370 hasConceptScore W4386072370C34974158 @default.
- W4386072370 hasConceptScore W4386072370C38652104 @default.
- W4386072370 hasConceptScore W4386072370C41008148 @default.
- W4386072370 hasConceptScore W4386072370C41608201 @default.
- W4386072370 hasConceptScore W4386072370C50644808 @default.
- W4386072370 hasLocation W43860723701 @default.
- W4386072370 hasOpenAccess W4386072370 @default.
- W4386072370 hasPrimaryLocation W43860723701 @default.
- W4386072370 hasRelatedWork W1978508564 @default.
- W4386072370 hasRelatedWork W2141506052 @default.
- W4386072370 hasRelatedWork W2279398222 @default.
- W4386072370 hasRelatedWork W2579318729 @default.
- W4386072370 hasRelatedWork W2791691546 @default.
- W4386072370 hasRelatedWork W2915754718 @default.
- W4386072370 hasRelatedWork W3082895349 @default.
- W4386072370 hasRelatedWork W4224919626 @default.
- W4386072370 hasRelatedWork W4299822940 @default.
- W4386072370 hasRelatedWork W4310220500 @default.
- W4386072370 isParatext "false" @default.
- W4386072370 isRetracted "false" @default.
- W4386072370 workType "article" @default.