Adversarial Representation Learning for Robust Privacy Preservation in Audio

Sound event detection systems are widely used in various applications such as surveillance and environmental monitoring where data is automatically collected, processed, and sent to a cloud for sound recognition.