Research map / Video / Video generation
Dual Reciprocal Learning of Language-based Human Motion Understanding and Generation
Chen Liang, Zhicheng Shi, Wenguan Wang, Yi Yang
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 DRL | TM2T*-T2M · HumanML3D T2M (Ori.) | FID: .092 -> .080 | Improved |
| add DRL | MotionGPT* · HumanML3D T2M (Ori.) | FID: .363 -> .327 | Improved |
| add DRL | TM2T*-T2M · HumanML3D T2M (Upd.) | FID: .063 -> .039 | Improved |
| add DRL | TM2T*-T2M · HumanML3D T2M (Ori.) | R-Precision Top-3: .742 -> .735 | Hurt |
| add DRL | TM2T*-M2T · HumanML3D M2T | CIDEr: 43.47 -> 50.90 | Improved |
| add DRL | MotionGPT* · HumanML3D M2T | CIDEr: 5.68 -> 13.26 | Improved |
| add DRL | TM2T*-M2T · HumanML3D M2T (Upd.) | R-Precision Top-1: .199 -> .210 | Improved |
| add unpaired motion data to dual learning (ablation) | DRL (TM2T*) · HumanML3D M2T | CIDEr: 47.28 -> 48.07 | Improved |
| add unpaired text data on top of unpaired motion (ablation) | DRL (TM2T*) · HumanML3D M2T | CIDEr: 48.07 -> 50.90 | Improved |
| add dual reciprocal learning using only paired data (ablation) | TM2T* baseline · HumanML3D T2M (Upd.) | FID: .063 -> .048 | Improved |
| increase unpaired motion from 3k to 12k clips (ablation) | DRL (TM2T*) · HumanML3D M2T | CIDEr: 45.02 -> 52.23 | Improved |
| reduce paired data to 1/8 with unpaired data (ablation) | DRL (TM2T*) · HumanML3D M2T | CIDEr: 43.47 -> 22.75 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- BAMM — BAMM: Bidirectional Autoregressive Motion Model (ECCV 2024) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MotionGPT: Human Motion as a Foreign Language (NeurIPS 2023 · similarity 0.90)
- VersatileMotion: A Unified Framework for Motion Synthesis and Comprehension (ECCV 2026 · similarity 0.89)
- A Unified Framework for Motion Reasoning and Generation in Human Interaction (ICCV 2025 · similarity 0.89)
- TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Texts (ECCV 2022 · similarity 0.89)
- The Language of Motion: Unifying Verbal and Non-verbal Language of 3D Human Motion (CVPR 2025 · similarity 0.88)
- ReMoGPT: Part-Level Retrieval-Augmented Motion-Language Models (AAAI 2025 · similarity 0.88)
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.
