Xingjian Du

transformer architecture (2) reinforcement learning (1) zero-shot learning (1) source separation (1) attention mechanism (1) weakly supervised learning (1) language modeling (1) multimodal learning (1) model architecture (1) efficient inference (1) speech enhancement (1) computational efficiency (1) audio source separation (1) audio-visual fusion (1) recurrent neural network (1) linear attention (1) cross-modal fusion (1) symbolic music generation (1) sound event detection (1) voice activity detection (1)