GhostVec: Directly Extracting Speaker Embedding from End-to-End Speech Recognition Model Using Adversarial Examples.

Xiaojiao Chen, Sheng Li, Hao Huang

Browse the full ICONIP paper archive.