Are We Ready for RL in Text-to-3D Generation? A Progressive Investigation

• URL: http://arxiv.org/abs/2512.10949v1
• Date: Thu, 11 Dec 2025 18:59:52 GMT
• Title: Are We Ready for RL in Text-to-3D Generation? A Progressive Investigation
• Authors: Yiwen Tang, Zoey Guo, Kaixin Zhu, Ray Zhang, Qizhi Chen, Dongzhi Jiang, Junli Liu, Bohan Zeng, Haoming Song, Delin Qu, Tianyi Bai, Dan Xu, Wentao Zhang, Bin Zhao,
• Abstract summary: Reinforcement learning (RL) has been successfully extended to enhance 2D image generation.<n>Applying RL to 3D generation remains largely unexplored due to the higher spatial complexity of 3D objects.<n>We conduct the first systematic study of RL for text-to-3D autoregressive generation across several dimensions.
• Score: 29.767530820783378
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Reinforcement learning (RL), earlier proven to be effective in large language and multi-modal models, has been successfully extended to enhance 2D image generation recently. However, applying RL to 3D generation remains largely unexplored due to the higher spatial complexity of 3D objects, which require globally consistent geometry and fine-grained local textures. This makes 3D generation significantly sensitive to reward designs and RL algorithms. To address these challenges, we conduct the first systematic study of RL for text-to-3D autoregressive generation across several dimensions. (1) Reward designs: We evaluate reward dimensions and model choices, showing that alignment with human preference is crucial, and that general multi-modal models provide robust signal for 3D attributes. (2) RL algorithms: We study GRPO variants, highlighting the effectiveness of token-level optimization, and further investigate the scaling of training data and iterations. (3) Text-to-3D Benchmarks: Since existing benchmarks fail to measure implicit reasoning abilities in 3D generation models, we introduce MME-3DR. (4) Advanced RL paradigms: Motivated by the natural hierarchy of 3D generation, we propose Hi-GRPO, which optimizes the global-to-local hierarchical 3D generation through dedicated reward ensembles. Based on these insights, we develop AR3D-R1, the first RL-enhanced text-to-3D model, expert from coarse shape to texture refinement. We hope this study provides insights into RL-driven reasoning for 3D generation. Code is released at https://github.com/Ivan-Tang-3D/3DGen-R1.

Related papers

• 3D and 4D World Modeling: A Survey [104.20852751473392]
  World modeling has become a cornerstone in AI research, enabling agents to understand, represent, and predict the dynamic environments they inhabit.<n>We introduce a structured taxonomy spanning video-based (VideoGen), occupancy-based (OccGen), and LiDAR-based (LiDARGen) approaches.<n>We discuss practical applications, identify open challenges, and highlight promising research directions.
  arXiv  Detail & Related papers  (2025-09-04T17:59:58Z)
• End-to-End Fine-Tuning of 3D Texture Generation using Differentiable Rewards [8.953379216683732]
  We propose an end-to-end differentiable, reinforcement-learning-free framework that embeds human feedback, expressed as differentiable reward functions, directly into the 3D texture pipeline.<n>By back-propagating preference signals through both geometric and appearance modules, our method generates textures that respect the 3D geometry structure and align with desired criteria.
  arXiv  Detail & Related papers  (2025-06-23T06:24:12Z)
• TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models [69.0220314849478]
  TripoSG is a new streamlined shape diffusion paradigm capable of generating high-fidelity 3D meshes with precise correspondence to input images.<n>The resulting 3D shapes exhibit enhanced detail due to high-resolution capabilities and demonstrate exceptional fidelity to input images.<n>To foster progress and innovation in the field of 3D generation, we will make our model publicly available.
  arXiv  Detail & Related papers  (2025-02-10T16:07:54Z)
• DIRECT-3D: Learning Direct Text-to-3D Generation on Massive Noisy 3D Data [50.164670363633704]
  We present DIRECT-3D, a diffusion-based 3D generative model for creating high-quality 3D assets from text prompts. Our model is directly trained on extensive noisy and unaligned in-the-wild' 3D assets. We achieve state-of-the-art performance in both single-class generation and text-to-3D generation.
  arXiv  Detail & Related papers  (2024-06-06T17:58:15Z)
• Retrieval-Augmented Score Distillation for Text-to-3D Generation [30.57225047257049]
  We introduce novel framework for retrieval-based quality enhancement in text-to-3D generation. We conduct extensive experiments to demonstrate that ReDream exhibits superior quality with increased geometric consistency.
  arXiv  Detail & Related papers  (2024-02-05T12:50:30Z)
• Sherpa3D: Boosting High-Fidelity Text-to-3D Generation via Coarse 3D Prior [52.44678180286886]
  2D diffusion models find a distillation approach that achieves excellent generalization and rich details without any 3D data. We propose Sherpa3D, a new text-to-3D framework that achieves high-fidelity, generalizability, and geometric consistency simultaneously.
  arXiv  Detail & Related papers  (2023-12-11T18:59:18Z)
• Text-to-3D Generation with Bidirectional Diffusion using both 2D and 3D priors [16.93758384693786]
  Bidirectional Diffusion(BiDiff) is a unified framework that incorporates both a 3D and a 2D diffusion process. Our model achieves high-quality, diverse, and scalable 3D generation.
  arXiv  Detail & Related papers  (2023-12-07T10:00:04Z)
• Lift3D: Synthesize 3D Training Data by Lifting 2D GAN to 3D Generative Radiance Field [16.15190186574068]
  We propose Lift3D, an inverted 2D-to-3D generation framework to achieve the data generation objectives. By lifting well-disentangled 2D GAN to 3D object NeRF, Lift3D provides explicit 3D information of generated objects. We evaluate the effectiveness of our framework by augmenting autonomous driving datasets.
  arXiv  Detail & Related papers  (2023-04-07T07:43:02Z)
• Get3DHuman: Lifting StyleGAN-Human into a 3D Generative Model using Pixel-aligned Reconstruction Priors [56.192682114114724]
  Get3DHuman is a novel 3D human framework that can significantly boost the realism and diversity of the generated outcomes. Our key observation is that the 3D generator can profit from human-related priors learned through 2D human generators and 3D reconstructors.
  arXiv  Detail & Related papers  (2023-02-02T15:37:46Z)
• Deep Generative Models on 3D Representations: A Survey [81.73385191402419]
  Generative models aim to learn the distribution of observed data by generating new instances. Recently, researchers started to shift focus from 2D to 3D space. representing 3D data poses significantly greater challenges.
  arXiv  Detail & Related papers  (2022-10-27T17:59:50Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.