What Was That Again? Certified Robustness for Automatic Speech Recognition
arXiv:2606.27698v1 Announce Type: new Abstract: Automatic Speech Recognition systems are notoriously both sensitive to adversarial and benign perturbations. While this has been repeatedly demonstrated using reference datasets, detecting such behaviors in deployed systems is incredibly challenging, due to the absence of oracle knowledge of the true transcription. We demonstrate that employing a certification-inspired mechanism can significantly decrease WER, increase recall, and decrease the Spea...
arXiv cs.LG
·Andrew C. Cullen, Neil Marchant, Jiani Xie, Paul Montague, Benjamin I. P. Rubinstein
·
// relacionados
Leia também
Blog
Verificando criptografia em Rust no SymCrypt, dos padrões ao código
Blog
Transplante de Tokenizador: Mitigando o Colapso Autorregressivo em ASR de Bengali Eficiente para Edge
Blog
Além dos Deslocamentos Temporais: Adaptando o Omni-LLM como um Avaliador Sem Referência para Modelos Generativos Audiovisuais
Blog