AINeutralarXiv – CS AI · Mar 276/10
🧠
TAAC: A gate into Trustable Audio Affective Computing
Researchers have developed TAAC, a framework for trustable audio-based depression diagnosis that protects user identity information while maintaining diagnostic accuracy. The system uses adversarial loss-based subspace decomposition to separate depression features from sensitive identity data, enabling secure AI-powered mental health screening.