mlsanityData to 7 Oct 2026

MoRL: Reinforced Reasoning for Unified Motion Understanding and Generation

arXiv 2602.14534 ↗First posted 16 Feb 2026cs.CVcited by 0

On the map: LLM reasoning → Spatial and embodied reasoning

Abstract (arXiv)

Human motion understanding and generation are crucial for vision and robotics but remain limited in reasoning capability and test-time planning. We propose MoRL, a unified multimodal motion model trained with supervised fine-tuning and reinforcement learning with verifiable rewards. Our task-specific reward design combines semantic alignment and reasoning coherence for understanding with physical plausibility and text-motion consistency for generation, improving both logical reasoning and perceptual realism. To further enhance inference, we introduce Chain-of-Motion (CoM), a test-time reasoning method that enables step-by-step planning and reflection. We also construct two large-scale CoT datasets, MoUnd-CoT-140K and MoGen-CoT-140K, to align motion sequences with reasoning traces and action descriptions. Experiments on HumanML3D and KIT-ML show that MoRL achieves significant gains over state-of-the-art baselines. Code: https://github.com/AIGeeksGroup/MoRL. Website: https://aigeeksgroup.github.io/MoRL.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning. Builds on 32 of its 54 references that are in the map. First posted on arXiv 16 Feb 2026.

Builds on

32 of 54 references.

References · 32

  1. 2019Expressive Body Capture: 3D Hands, Face, and Body from a Single Imagecited by 1,600
  2. 2022Human Motion Diffusion Modelcited by 1,062
  3. 2022MotionDiffuse: Text-Driven Human Motion Generation with Diffusion Modelcited by 768
  4. 2023MotionGPT: Human Motion as a Foreign Languagecited by 521
  5. 2025Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Modelcited by 450
  6. 2022TM2T: Stochastic and Tokenized Modeling for the Reciprocal Generation of 3D Human Motions and Textscited by 357
  7. 2016The KIT Motion-Language Datasetcited by 334
  8. 2019Language2Pose: Natural Language Grounded Pose Forecastingcited by 257
  9. 2023MotionGPT: Finetuned LLMs Are General-Purpose Motion Generatorscited by 145
  10. 2024Marco-o1: Towards Open Reasoning Models for Open-Ended Solutionscited by 142
  11. 2021Text2Gestures: A Transformer-Based Network for Generating Emotive Body Gestures for Virtual Agentscited by 106
  12. 2023FineMoGen: Fine-Grained Spatio-Temporal Motion Generation and Editingcited by 105
20 more
  1. 2024MotionLLM: Understanding Human Behaviors from Human Motions and Videoscited by 89
  2. 2023AvatarGPT: All-in-One Framework for Motion Understanding, Planning, Generation and Beyondcited by 63
  3. 2017Learning a bidirectional mapping between human whole-body motion and natural language using deep recurrent neural networkscited by 61
  4. 2024LaMP: Language-Motion Pretraining for Motion Generation, Retrieval, and Captioningcited by 58
  5. 2024MotionGPT-2: A General-Purpose Motion-Language Model for Motion Generation and Understandingcited by 58
  6. 2024Motion-Agent: A Conversational Framework for Human Motion Generation with LLMscited by 46
  7. 2025Motion Anything: Any to Motion Generationcited by 40
  8. 2024TextCoT: Zoom In for Enhanced Multimodal Text-Rich Image Understandingcited by 37
  9. 2024MotionChain: Conversational Motion Controllers via Multimodal Promptscited by 30
  10. 2024InfiniMotion: Mamba Boosts Memory in Transformer for Arbitrary Long Motion Generationcited by 23
  11. 2024Motion Avatar: Generate Human and Animal Avatars with Arbitrary Motioncited by 19
  12. 2024CoMA: Compositional Human Motion Generation with Multi-modal Agentscited by 18
  13. 2024Aligning Human Motion Generation with Human Perceptionscited by 16
  14. 2024KMM: Key Frame Mask Mamba for Extended Motion Generationcited by 16
  15. 2025ReMoMask: Retrieval-Augmented Masked Motion Generationcited by 11
  16. 2023T2M-HiFiGPT: Generating High Quality Human Motion from Textual Descriptions with Residual Discrete Representationscited by 10
  17. 2024MotionRL: Align Text-to-Motion Generation to Human Preferences with Multi-Reward Reinforcement Learningcited by 10
  18. 2026SafeMo: Linguistically Grounded Unlearning for Trustworthy Text-to-Motion Generationcited by 6
  19. 2023Guided Attention for Interpretable Motion Captioningcited by 5
  20. 2023Motion2Language, unsupervised learning of synchronized semantic motion segmentationcited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Motion-R1: Enhancing Motion Generation with Decomposed Chain-of-Thought and RL Binding
  2. 2026UniMo: Unified Motion Generation and Understanding with Chain of Thought
  3. 2026MotionMERGE: A Multi-granular Framework for Human Motion Editing, Reasoning, Generation, and Explanation
  4. 2026Motion-as-Prompt: Enhancing Motion Reasoning in Multimodal Large Language Models via Motion-Guided Cross-Frame Visual Prompting
  5. 2025Unveiling Chain of Step Reasoning for Vision-Language Models with Fine-grained Rewards
  6. 2026ThinkingVLA: Interleaved Vision and Language Reasoning for Robotic Manipulation
  7. 2025Do What You Say: Steering Vision-Language-Action Models via Runtime Reasoning-Action Alignment Verification
  8. 2026Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation