Research map / Video / Video generation
M2D2M: Multi-Motion Generation from Text with Discrete Diffusion Models
Seunggeun Chi, Hyung-gun Chi, Hengbo Ma, Nakul Agarwal, Faizan Siddiqui, Karthik Ramani, Kwonjoon Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use dynamic transition probability beta(t,d) instead of static beta_t (ablation) | M2D2M · HumanML3D single-motion | FID: 0.115 -> 0.087 | Improved |
| use dynamic transition probability beta(t,d) (ablation) | M2D2M · KIT-ML single-motion | FID: 0.404 -> 0.378 | Improved |
| use Two-Phase Sampling instead of Handshake (ablation) | VQ-Diffusion multi-motion · HumanML3D multi-motion | individual FID: 1.279 -> 0.254 | Improved |
| use Two-Phase Sampling instead of SLERP (ablation) | VQ-Diffusion multi-motion · HumanML3D multi-motion | individual FID: 1.402 -> 0.254 | Improved |
| reduce joint sampling steps Ts from 100 to 90 (with dynamic probability) (ablation) | M2D2M · HumanML3D multi-motion | individual FID: FID 0.196 -> 0.253; transition Jerk 1.248 -> 1.238 | Mixed |
| M2D2M vs T2M-GPT on KIT-ML transition quality | M2D2M · KIT-ML multi-motion transition | FID: T2M-GPT 14.494 vs M2D2M 15.843 | Hurt |
| M2D2M vs M2DM on single-motion diversity of outputs | M2D2M · HumanML3D single-motion | MModality: 3.587 vs 2.115 | Hurt |
Closest papers by meaning
- AMD: Autoregressive Motion Diffusion (AAAI 2024 · similarity 0.92)
- Rethinking Diffusion for Text-Driven Human Motion Generation: Redundant Representations, Evaluation, and Masked Autoregression (CVPR 2025 · similarity 0.91)
- Towards Detailed Text-to-Motion Synthesis via Basic-to-Advanced Hierarchical Diffusion Model (AAAI 2024 · similarity 0.90)
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.90)
- AMD: Anatomical Motion Diffusion with Interpretable Motion Decomposition and Fusion (AAAI 2024 · similarity 0.89)
- MotionMix: Weakly-Supervised Diffusion for Controllable Motion Generation (AAAI 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
