PZ
Pengcheng Zhang
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
2 records found
1
Deep neural networks (DNNs) have been widely and successfully adopted and deployed in various applications of speech recognition. Recently, a few works revealed that these models are vulnerable to backdoor attacks, where the adversaries can implant malicious prediction behaviors into victim models by poisoning their training process. In this paper, we revisit poison-only backdoor attacks against speech recognition. We reveal that existing methods are not stealthy since their trigger patterns are perceptible to humans or machine detection. This limitation is mostly because their trigger patterns are simple noises or separable and distinctive clips. Motivated by these findings, we propose to exploit elements of sound ( e.g ., pitch and timbre) to design more stealthy yet effective poison-only backdoor attacks. Specifically, we insert a short-duration high-pitched signal as the trigger and increase the pitch of remaining audio clips to ‘mask’ it for designing stealthy pitch-based triggers. We manipulate timbre features of victim audio to design the stealthy timbre-based attack and design a voiceprint selection module to facilitate the multi-backdoor attack. Our attacks can generate more ‘natural’ poisoned samples and therefore are more stealthy. Extensive experiments are conducted on benchmark datasets, which verify the effectiveness of our attacks under different settings ( e.g ., all-to-one, all-to-all, clean-label, physical, and multi-backdoor settings) and their stealthiness. Our methods achieve attack success rates of over 95% in most cases and are nearly undetectable. The code for reproducing main experiments are available at https://github.com/HanboCai/BadSpeech_SoE .
...
Deep neural networks (DNNs) have been widely and successfully adopted and deployed in various applications of speech recognition. Recently, a few works revealed that these models are vulnerable to backdoor attacks, where the adversaries can implant malicious prediction behaviors into victim models by poisoning their training process. In this paper, we revisit poison-only backdoor attacks against speech recognition. We reveal that existing methods are not stealthy since their trigger patterns are perceptible to humans or machine detection. This limitation is mostly because their trigger patterns are simple noises or separable and distinctive clips. Motivated by these findings, we propose to exploit elements of sound ( e.g ., pitch and timbre) to design more stealthy yet effective poison-only backdoor attacks. Specifically, we insert a short-duration high-pitched signal as the trigger and increase the pitch of remaining audio clips to ‘mask’ it for designing stealthy pitch-based triggers. We manipulate timbre features of victim audio to design the stealthy timbre-based attack and design a voiceprint selection module to facilitate the multi-backdoor attack. Our attacks can generate more ‘natural’ poisoned samples and therefore are more stealthy. Extensive experiments are conducted on benchmark datasets, which verify the effectiveness of our attacks under different settings ( e.g ., all-to-one, all-to-all, clean-label, physical, and multi-backdoor settings) and their stealthiness. Our methods achieve attack success rates of over 95% in most cases and are nearly undetectable. The code for reproducing main experiments are available at https://github.com/HanboCai/BadSpeech_SoE .
Cockcroft-Walton voltage multiplier circuit is widely used for high voltage generation circuit. The diode reverse recovery effect of Cockcroft-Walton voltage multiplier circuit is investigated by analysis and circuit simulation. According to the analysis and circuit simulation study, it can be concluded that the multiplier diode reverse recovery problem is mainly caused by the diodes in the first stage voltage multiplier. The most effective and economic way to alleviate the diode reverse recovery problem is employing diodes with good reverse recovery performance such as silicon carbide Schottky diodes only in the first stage for good system performance. The experimental results of 3 stages Cockcroft-Walton voltage multiplier circuit hardware prototype operating at 300kHz switching frequency validate the concept based on analysis and simulation study. The silicon carbide diode without reverse recovery needs to be used only in the first stage of voltage multiplier circuit to effectively mitigate the reverse recovery problems in high frequency operations with good circuit performance.
...
Cockcroft-Walton voltage multiplier circuit is widely used for high voltage generation circuit. The diode reverse recovery effect of Cockcroft-Walton voltage multiplier circuit is investigated by analysis and circuit simulation. According to the analysis and circuit simulation study, it can be concluded that the multiplier diode reverse recovery problem is mainly caused by the diodes in the first stage voltage multiplier. The most effective and economic way to alleviate the diode reverse recovery problem is employing diodes with good reverse recovery performance such as silicon carbide Schottky diodes only in the first stage for good system performance. The experimental results of 3 stages Cockcroft-Walton voltage multiplier circuit hardware prototype operating at 300kHz switching frequency validate the concept based on analysis and simulation study. The silicon carbide diode without reverse recovery needs to be used only in the first stage of voltage multiplier circuit to effectively mitigate the reverse recovery problems in high frequency operations with good circuit performance.