Sonic-AMG: Streaming Human Motion Generation from Spatial Audio
cs.CV, cs.GR, cs.LG, cs.RO, cs.SD
Submitted: 2026-10-09
Updated: 2026-10-09
Project page: https://jingchengni.com/sonic-amg
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- GestureDiffuCLIP: Gesture Diffusion Model with CLIP Latents
- Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion
- SoundSpaces: Audio-Visual Navigation in 3D Environments
- Semantic Audio-Visual Navigation
- FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders
- Self Forcing: Bridging the Train-Test Gap in Autoregressive Video Diffusion
- DiscoForcing: A Unified Framework for Real-Time Audio-Driven Character Control with Diffusion Forcing
- MotionGPT: Human Motion as a Foreign Language
- Autoregressive Image Generation without Vector Quantization
- Flow Matching for Generative Modeling
- GestureFAR: Streaming Co-Speech Gesture Generation with Flow Autoregression
- Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
- SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control
- EDGE: Editable Dance Generation From Music
- MotionStreamer: Streaming Motion Generation via Diffusion-based Autoregressive Model in Causal Latent Space
- MOSPA: Human Motion Generation Driven by Spatial Audio
- MoSAT: Human Motion Generation from Spatial Audio and Textual Description
- Speech Gesture Generation from the Trimodal Context of Text, Audio, and Speaker Identity
- DartControl: A Diffusion-Based Autoregressive Motion Model for Real-Time Text-Driven Motion Control
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models