Zhiyong Wu

From Inpainting to Editing: A Self-Bootstrapping Paradigm for Context-Rich Visual Dubbing

Hierarchical Semantic-Acoustic Modeling via Semi-Discrete Residual Representations for Expressive End-to-End Speech Synthesis

Self-Guidance: Training VQ-VAE Decoders to be Robust to Quantization Artifacts for High-Fidelity Neural Speech Codec

WebRAGent: Retrieval-Augmented Generation for Multimodal Web Agent Planning

AlignMark: Content-Aligned Audio Watermarking for Robustness Against Neural Transformations

LeVo: High-Quality Song Generation with Multi-Preference Alignment

E2E-VGuard: Adversarial Prevention for Production LLM-based End-To-End Speech Synthesis

RFWave: Multi-band Rectified Flow for Audio Waveform Reconstruction