Feng Deng

multi-task learning (3) audio source separation (3) signal-to-distortion ratio (2) channel attention (1) neural architecture search (1) visual grounding (1) flow matching (1) speech enhancement (1) latent diffusion model (1) window attention (1) audio captioning (1) model efficiency (1) noise estimation (1) residual learning (1) text-to-audio generation (1) audio synthesis (1) keyword prediction (1) spatial attention (1) multi-task loss (1) cnn transformer (1)