Matches in SemOpenAlex for { <https://semopenalex.org/work/W3138128528> ?p ?o ?g. }
Showing items 1 to 96 of
96
with 100 items per page.
- W3138128528 endingPage "2058" @default.
- W3138128528 startingPage "2044" @default.
- W3138128528 abstract "This paper introduces a defense approach against end-to-end adversarial attacks developed for cutting-edge speech-to-text systems. The proposed defense algorithm has four steps. First, we use the short-time Fourier transform to represent speech signals with 2D spectrograms. Second, we iteratively find a safe vector using a spectrogram subspace projection operation. This operation minimizes the chordal distance adjustment between spectrograms with an additional regularization term. Third, we synthesize a spectrogram with such a safe vector using a novel GAN architecture trained with Sobolev integral probability metric. We impose an additional constraint on the generator network to improve the model’s performance in terms of stability and the total number of learned modes. Finally, we reconstruct the signal from the synthesized spectrogram and the Griffin-Lim phase approximation technique. We evaluate the proposed defense approach against six strong white and black-box adversarial attacks on DeepSpeech, Kaldi, and Lingvo models. The experimental results show that our algorithm outperforms other state-of-the-art defense algorithms in terms of accuracy and signal quality." @default.
- W3138128528 created "2021-03-29" @default.
- W3138128528 creator A5051056084 @default.
- W3138128528 creator A5074262951 @default.
- W3138128528 creator A5085673698 @default.
- W3138128528 date "2022-01-01" @default.
- W3138128528 modified "2023-10-16" @default.
- W3138128528 title "Multidiscriminator Sobolev Defense-GAN Against Adversarial Attacks for End-to-End Speech Systems" @default.
- W3138128528 cites W1269046860 @default.
- W3138128528 cites W1494198834 @default.
- W3138128528 cites W1552314771 @default.
- W3138128528 cites W2000992741 @default.
- W3138128528 cites W2030680983 @default.
- W3138128528 cites W2038484192 @default.
- W3138128528 cites W2052666245 @default.
- W3138128528 cites W2112739286 @default.
- W3138128528 cites W2120847449 @default.
- W3138128528 cites W2127141656 @default.
- W3138128528 cites W2132196015 @default.
- W3138128528 cites W2141998673 @default.
- W3138128528 cites W2143612262 @default.
- W3138128528 cites W2158185319 @default.
- W3138128528 cites W2194775991 @default.
- W3138128528 cites W2328111639 @default.
- W3138128528 cites W2768083292 @default.
- W3138128528 cites W2898435086 @default.
- W3138128528 cites W2899730059 @default.
- W3138128528 cites W2940177920 @default.
- W3138128528 cites W2962765321 @default.
- W3138128528 cites W2962826786 @default.
- W3138128528 cites W2962911378 @default.
- W3138128528 cites W2963057973 @default.
- W3138128528 cites W2963077926 @default.
- W3138128528 cites W2964301649 @default.
- W3138128528 cites W2973057252 @default.
- W3138128528 cites W2973252307 @default.
- W3138128528 cites W2983148138 @default.
- W3138128528 cites W2993396030 @default.
- W3138128528 cites W3006816054 @default.
- W3138128528 cites W3016206375 @default.
- W3138128528 cites W3034222740 @default.
- W3138128528 cites W3095722698 @default.
- W3138128528 cites W3095753917 @default.
- W3138128528 cites W3096366160 @default.
- W3138128528 cites W3110715780 @default.
- W3138128528 cites W3125709657 @default.
- W3138128528 cites W3157058436 @default.
- W3138128528 cites W3163619259 @default.
- W3138128528 cites W4232648653 @default.
- W3138128528 cites W4300672471 @default.
- W3138128528 doi "https://doi.org/10.1109/tifs.2022.3175603" @default.
- W3138128528 hasPublicationYear "2022" @default.
- W3138128528 type Work @default.
- W3138128528 sameAs 3138128528 @default.
- W3138128528 citedByCount "7" @default.
- W3138128528 countsByYear W31381285282022 @default.
- W3138128528 countsByYear W31381285282023 @default.
- W3138128528 crossrefType "journal-article" @default.
- W3138128528 hasAuthorship W3138128528A5051056084 @default.
- W3138128528 hasAuthorship W3138128528A5074262951 @default.
- W3138128528 hasAuthorship W3138128528A5085673698 @default.
- W3138128528 hasBestOaLocation W31381285282 @default.
- W3138128528 hasConcept C11413529 @default.
- W3138128528 hasConcept C154945302 @default.
- W3138128528 hasConcept C28490314 @default.
- W3138128528 hasConcept C32834561 @default.
- W3138128528 hasConcept C41008148 @default.
- W3138128528 hasConcept C45273575 @default.
- W3138128528 hasConceptScore W3138128528C11413529 @default.
- W3138128528 hasConceptScore W3138128528C154945302 @default.
- W3138128528 hasConceptScore W3138128528C28490314 @default.
- W3138128528 hasConceptScore W3138128528C32834561 @default.
- W3138128528 hasConceptScore W3138128528C41008148 @default.
- W3138128528 hasConceptScore W3138128528C45273575 @default.
- W3138128528 hasFunder F4320334593 @default.
- W3138128528 hasLocation W31381285281 @default.
- W3138128528 hasLocation W31381285282 @default.
- W3138128528 hasOpenAccess W3138128528 @default.
- W3138128528 hasPrimaryLocation W31381285281 @default.
- W3138128528 hasRelatedWork W2065296656 @default.
- W3138128528 hasRelatedWork W2138997758 @default.
- W3138128528 hasRelatedWork W2395582662 @default.
- W3138128528 hasRelatedWork W2766960583 @default.
- W3138128528 hasRelatedWork W2897924318 @default.
- W3138128528 hasRelatedWork W2900550435 @default.
- W3138128528 hasRelatedWork W2973062255 @default.
- W3138128528 hasRelatedWork W3133517635 @default.
- W3138128528 hasRelatedWork W3195104037 @default.
- W3138128528 hasRelatedWork W4301857073 @default.
- W3138128528 hasVolume "17" @default.
- W3138128528 isParatext "false" @default.
- W3138128528 isRetracted "false" @default.
- W3138128528 magId "3138128528" @default.
- W3138128528 workType "article" @default.