all AI news
DistriBlock: Identifying adversarial audio samples by leveraging characteristics of the output distribution
Feb. 16, 2024, 5:44 a.m. | Mat\'ias P. Pizarro B., Dorothea Kolossa, Asja Fischer
cs.LG updates on arXiv.org arxiv.org
Abstract: Adversarial attacks can mislead automatic speech recognition (ASR) systems into predicting an arbitrary target text, thus posing a clear security threat. To prevent such attacks, we propose DistriBlock, an efficient detection strategy applicable to any ASR system that predicts a probability distribution over output tokens in each time step. We measure a set of characteristics of this distribution: the median, maximum, and minimum over the output probabilities, the entropy of the distribution, as well as …
abstract adversarial adversarial attacks arxiv asr attacks audio automatic speech recognition clear cs.cr cs.lg cs.sd detection distribution eess.as probability recognition samples security speech speech recognition strategy systems text threat type
More from arxiv.org / cs.LG updates on arXiv.org
Jobs in AI, ML, Big Data
AI Research Scientist
@ Vara | Berlin, Germany and Remote
Data Architect
@ University of Texas at Austin | Austin, TX
Data ETL Engineer
@ University of Texas at Austin | Austin, TX
Lead GNSS Data Scientist
@ Lurra Systems | Melbourne
Senior Machine Learning Engineer (MLOps)
@ Promaton | Remote, Europe
Data Analyst (Digital Business Analyst)
@ Activate Interactive Pte Ltd | Singapore, Central Singapore, Singapore