A speech backdoor attack uses SLLM-generated timbre and emotion triggers with MGDA-balanced training to implant multiple effective backdoors in speech classifiers.
A backdoor attack against lstm- based text classification systems,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SPBA: Utilizing Speech Large Language Model for Backdoor Attacks on Speech Classification Models
A speech backdoor attack uses SLLM-generated timbre and emotion triggers with MGDA-balanced training to implant multiple effective backdoors in speech classifiers.