Research map / Image generation / Text-to-image diffusion
Being Comes From Not-Being: Open-Vocabulary Text-to-Motion Generation With Wordless Training
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 OOHMG instead of MotionCLIP | Text-to-motion generation · BABEL | Top50: 0.0661 -> 0.6494 | Helped |
| Use OOHMG instead of AvatarCLIP | Text-to-motion generation · BABEL | In-distrib.: 0.0407 -> 0.0205 | Helped |
| Use wordless training with random text features rather than real texts (ablation) | Text-to-pose generator · BABEL | Top50: 0.2104 -> 0.7224 | Helped |
| Use wordless training with random text features rather than real texts (ablation) | Text-to-pose generator · BABEL | In-distrib.: 0.1210 -> 0.0111 | Helped |
| Use TPA contrastive loss instead of maximizing the TPA score (ablation) | Text-to-pose generator trained on real text · BABEL | Top50: 0.2090 -> 0.2104 | No clear effect |
| Add L2 regularization on predicted latent poses (ablation) | Wordless text-to-pose generator · BABEL | In-distrib.: 0.0111 -> 0.0015 | Helped |
| Add an initial pose to the motion prompt (ablation) | OOHMG · BABEL text-to-motion | Top50: 0.6154 -> 0.6494 | Helped |
| Optimize text-to-pose generation using TPA score maximization rather than L_TPA (ablation) | Text-to-pose generator · BABEL | CLIP Score: 0.2491 vs 0.2557; 0.2524 vs 0.2613; 0.2620 vs 0.2601 | Mixed |
| Increase L2 regularization strength from 0 to 0.15 (ablation) | Wordless text-to-pose generator · BABEL | Top50: 0.7224 -> 0.6446 | Hurt |
| Increase L2 regularization strength from 0 to 0.15 (ablation) | Wordless text-to-pose generator · BABEL | In-distrib.: 0.0111 -> 0.0012 | Helped |
Closest papers by meaning
- SnapMoGen: Human Motion Generation from Expressive Texts (NeurIPS 2025 · similarity 0.89)
- OMG: Towards Open-vocabulary Motion Generation via Mixture of Controllers (CVPR 2024 · similarity 0.89)
- MMM: Generative Masked Motion Model (CVPR 2024 · similarity 0.88)
- AToM: Aligning Text-to-Motion Model at Event-Level with GPT-4Vision Reward (CVPR 2025 · similarity 0.87)
- Enhancing Motion in Text-to-Video Generation with Decomposed Encoding and Conditioning (NeurIPS 2024 · similarity 0.87)
- Generating Human Motion in 3D Scenes from Text Descriptions (CVPR 2024 · 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.