2 papers
cs.LG2025
Data Curation Matters: Model Collapse and Spurious Shift Performance Prediction from Training on Uncurated Text Embeddings
Lucas Mattioli, Youness Ait Hadichou, Sabrina Chaouche +1
Training models on uncurated Text Embeddings (TEs) derived from raw tabular data can lead to a severe failure mode known as model collapse, where predictions converge to a single c…
eess.AS2024
Detecting and Defending Against Adversarial Attacks on Automatic Speech Recognition via Diffusion Models
Nikolai L. Kühne, Astrid H. F. Kitchen, Marie S. Jensen +4
Automatic speech recognition (ASR) systems are known to be vulnerable to adversarial attacks. This paper addresses detection and defence against targeted white-box attacks on speec…