Research map / Video / Video generation
ReMoMask: Retrieval-Augmented Masked Motion Generation
Zhengdao Li, siheng wang, Zeyu Zhang, Hao Tang
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 |
|---|---|---|---|
| add bidirectional momentum model (BMM) retriever (ablation) | ReMoMask · HumanML3D | R-Precision Top1: 0.445 -> 0.531 | Improved |
| add bidirectional momentum model (BMM) retriever (ablation) | ReMoMask · HumanML3D | FID: 0.825 -> 0.411 | Improved |
| add SSTA instead of feature concatenation (ablation) | ReMoMask · HumanML3D | MultiModality: 1.094 -> 2.823 | Improved |
| add SSTA instead of feature concatenation (ablation) | ReMoMask · HumanML3D | MM Dist: 3.04 -> 2.865 | Improved |
| add RAG classifier-free guidance (ablation) | ReMoMask · HumanML3D | R-Precision Top1: 0.411 -> 0.531 | Improved |
| use local retrieval instead of global retrieval (ablation) | ReMoMask · HumanML3D | R-Precision Top3: 0.733 -> 0.813 | Improved |
| add text queue to motion queue (ablation) | BMM retriever · HumanML3D text-to-motion retrieval | R1: 12.44 -> 13.76 | Improved |
| add motion queue only, without text queue (ablation) | BMM retriever · HumanML3D motion-to-text retrieval | R1: 10.50 -> 0.70 | Hurt |
| add bidirectional queues (ablation) | BMM retriever · HumanML3D motion-to-text retrieval | R1: 10.50 -> 14.80 | Improved |
| use BMM retriever instead of PL-TMR | retriever · HumanML3D motion-to-text retrieval | R10: 39.11 -> 34.61 | Hurt |
| use ReMoMask instead of MoMask | text-to-motion model · HumanML3D | R-Precision Top1: 0.521 -> 0.531 | Improved |
| use ReMoMask instead of MoGenTS | text-to-motion model · HumanML3D | FID: 0.033 -> 0.099 | Hurt |
Closest papers by meaning
- HGM³: Hierarchical Generative Masked Motion Modeling with Hard Token Mining (ICLR 2025 · similarity 0.91)
- ReMoGPT: Part-Level Retrieval-Augmented Motion-Language Models (AAAI 2025 · similarity 0.90)
- MMM: Generative Masked Motion Model (CVPR 2024 · similarity 0.89)
- VimoRAG: Video-based Retrieval-augmented 3D Motion Generation for Motion Language Models (NeurIPS 2025 · similarity 0.89)
- SnapMoGen: Human Motion Generation from Expressive Texts (NeurIPS 2025 · similarity 0.87)
- MotionRAG: Motion Retrieval-Augmented Image-to-Video Generation (NeurIPS 2025 · similarity 0.87)
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.
