5 papers
DRL-CLBA: A Clean Label Backdoor Attack for Speech Classification via DDPG Reinforcement Learning
Yueming Huang, Wenhan Yao, Fen Xiao +2
Deep learning models for speech classification are vulnerable to backdoor attacks, where malicious triggers cause misclassification at inference time. While sample-specific attacks…
Pmeta-TLA: Backdoor Attacks for Speech Classification Models via Meta-Learning with Timbre Leakage Attack
Yueming Huang, Wenhan Yao, Fen Xiao +2
Recently, speech classification methods have gained widespread adoption in intelligent gadgets. Current study indicates that backdoor attacks provide a substantial security concern…
Pureformer-VC: Non-parallel Voice Conversion with Pure Stylized Transformer Blocks and Triplet Discriminative Training
Wenhan Yao, Fen Xiao, Xiarun Chen +3
As a foundational technology for intelligent human-computer interaction, voice conversion (VC) seeks to transform speech from any source timbre into any target timbre. Traditional…
SPBA: Utilizing Speech Large Language Model for Backdoor Attacks on Speech Classification Models
Wenhan Yao, Fen Xiao, Xiarun Chen +3
Deep speech classification tasks, including keyword spotting and speaker verification, are vital in speech-based human-computer interaction. Recently, the security of these technol…
Pureformer-VC: Non-parallel One-Shot Voice Conversion with Pure Transformer Blocks and Triplet Discriminative Training
Wenhan Yao, Zedong Xing, Xiarun Chen +3
One-shot voice conversion(VC) aims to change the timbre of any source speech to match that of the target speaker with only one speech sample. Existing style transfer-based VC metho…