AASIST
PulseAugur coverage of AASIST — every cluster mentioning AASIST across labs, papers, and developer communities, ranked by signal.
-
Ensemble deep learning system improves AI-altered video detection
Researchers have developed an ensemble deep learning system to detect AI-altered videos by combining audio and visual analysis. The system utilizes AASIST for audio detection and EfficientNet, XceptionNet, and MesoNet f…
-
New Browser Plugin Offers Lightweight, On-Device Audio Deepfake Detection
Researchers have developed a lightweight, on-device model for detecting audio deepfakes, designed to address privacy concerns associated with cloud-based solutions. This new model, integrated into a browser plugin, util…
-
New SpAArSIST model enhances audio anti-spoofing efficiency
Researchers have developed SpAArSIST, an optimized version of the AASIST model for anti-spoofing in audio. This new configuration reduces computational requirements by over 20% and model size by 4%, while significantly …
-
New method explains deepfake speech detector decisions
Researchers have developed a new method to understand how deepfake speech detectors make their decisions. By using Integrated Gradients on self-supervised representations, the technique can pinpoint specific moments in …