Beyond Noise Steering: Dual-Latent Space Reinforcement Learning for Generative Robot Policy

• URL: http://arxiv.org/abs/2609.11270v1
• Date: Thu, 10 Sep 2026 09:04:04 GMT
• Title: Beyond Noise Steering: Dual-Latent Space Reinforcement Learning for Generative Robot Policy
• Abstract summary: Pretrained generative robot policies learn expressive action prior to demonstrations.<n>Existing reinforcement learning methods only steer the noisy space but fail to modulate intermediate action representations during the generation process.<n>We propose a novel Dual-Latent Space Reinforcement Learning framework, which complements initial-noise steering with representation-level control inside the frozen generator.
• Score: 13.530890681678619
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Pretrained generative robot policies learn expressive action priors from demonstrations. However, existing reinforcement learning methods only steer the noisy space but fail to modulate intermediate action representations during the generation process, resulting in performance degradation and inefficiency. To address this limitation, we propose a novel Dual-Latent Space Reinforcement Learning (DLSRL) framework, which complements initial-noise steering with representation-level control inside the frozen generator. Specifically, our actor network predicts two distinct latent variables: an initial-noise latent variable that steers behavior generation, and an action-representation latent variable for intermediate feature modulation. Moreover, this representation latent variable is mapped to adapter features and ingeniously injected into the hidden states of intermediate action tokens via residual connections. Our dual-control design enables direct adjustment of action representations without updating the base policy. Experiments across generative policy architectures and robotic manipulation tasks show that DLSRL effectively accelerates online robot policy adaptation and achieves competitive performance. Our code is available at \href{https://github.com/xianchaoxiu/DLSRL}{https://github.com/xianchaoxiu/DLSRL}.

Related papers

• StructRL: Structured Action-Space Exploration for Flow-Based VLAs [81.43458629514139]
  Flow-based Vision-Language-Action models are now widely used for continuous robotic manipulation.<n>Online reinforcement learning is emerging as a key technique for adapting them to new tasks.<n>We show that effective robot exploration calls for structured noise: temporally smooth and scaled differently across action groups.
  arXiv  Detail & Related papers  (2026-08-15T09:31:29Z)
• SLIM-0.5B: Learning Action-Grounded Predictive Latents for Robot Manipulation [35.54319311306262]
  We present SLIM, a compact latent interaction policy.<n>It learns representations through self-supervised masked trajectory prediction.<n>SLIM matches or exceeds representative large-scale VLA and world-action-model baselines with fewer parameters.
  arXiv  Detail & Related papers  (2026-08-10T15:58:39Z)
• FlowDAgger: Human-in-the-Loop Adaptation of Generative Robot Policies in Latent Space [15.351960740594201]
  FlowDAgger is a sample- and compute-efficient method for adapting frozen generative robot policies from human interventions in latent space.<n>Our key idea is action inversion: each human expert action is mapped to the noise that would have produced it under the frozen base policy.<n>We evaluate FlowDAgger in simulation and on real-world bimanual and single-arm manipulation.
  arXiv  Detail & Related papers  (2026-07-09T19:07:33Z)
• Learning Object Manipulation from Scratch via Contrastive Interaction [29.976379315147756]
  Contrastive Reinforcement Learning (CRL) has seen recent success in a wide variety of goal-conditioned robotics tasks by learning structured representations of the dynamics.<n>We argue that a key source of this difficulty is object-centric interaction, such as contact or grasping, that induces distinct changes in the underlying dynamic modes.<n>In this work, we formulate manipulation dynamics as a piecewise-smooth Markov process and show that interaction-induced mode changes create piecewise nonlinear reachability structures.
  arXiv  Detail & Related papers  (2026-06-10T00:06:24Z)
• Primary-Fine Decoupling for Action Generation in Robotic Imitation [91.2899765310853]
  Multi-modal distribution in robotic manipulation action sequences poses critical challenges for imitation learning.<n>We propose Primary-Fine Decoupling for Action Generation (PF-DAG), a two-stage framework that decouples coarse action consistency from fine-grained variations.<n>PF-DAG outperforms state-of-the-art baselines across 56 tasks from Adroit, DexArt, and MetaWorld benchmarks.
  arXiv  Detail & Related papers  (2026-02-25T08:36:45Z)
• FreeAction: Training-Free Techniques for Enhanced Fidelity of Trajectory-to-Video Generation [50.39748673817223]
  We introduce two training-free, inference-time techniques that fully exploit explicit action parameters in robot video generation.<n>First, action-scaled classifier-free guidance dynamically modulates guidance strength in proportion to action magnitude, enhancing controllability over motion intensity.<n>Second, action-scaled noise truncation adjusts the distribution of initially sampled noise to better align with the desired motion dynamics.
  arXiv  Detail & Related papers  (2025-09-29T03:30:40Z)
• One-Step Diffusion Policy: Fast Visuomotor Policies via Diffusion Distillation [80.71541671907426]
  OneStep Diffusion Policy (OneDP) is a novel approach that distills knowledge from pre-trained diffusion policies into a single-step action generator. OneDP significantly accelerates response times for robotic control tasks.
  arXiv  Detail & Related papers  (2024-10-28T17:54:31Z)
• Autoregressive Action Sequence Learning for Robotic Manipulation [32.9580007141312]
  Existing autoregressive architectures generate end-effector waypoints sequentially as word tokens in language modeling.<n>We extend causal transformers' single-token prediction to support predicting a variable number of tokens in a single step.<n>We propose the Autoregressive Policy architecture, which solves manipulation tasks by generating hybrid action sequences.
  arXiv  Detail & Related papers  (2024-10-04T04:07:15Z)
• Training and Evaluation of Deep Policies using Reinforcement Learning and Generative Models [67.78935378952146]
  GenRL is a framework for solving sequential decision-making problems. It exploits the combination of reinforcement learning and latent variable generative models. We experimentally determine the characteristics of generative models that have most influence on the performance of the final policy training.
  arXiv  Detail & Related papers  (2022-04-18T22:02:32Z)
• Neural Dynamic Policies for End-to-End Sensorimotor Learning [51.24542903398335]
  The current dominant paradigm in sensorimotor control, whether imitation or reinforcement learning, is to train policies directly in raw action spaces. We propose Neural Dynamic Policies (NDPs) that make predictions in trajectory distribution space. NDPs outperform the prior state-of-the-art in terms of either efficiency or performance across several robotic control tasks.
  arXiv  Detail & Related papers  (2020-12-04T18:59:32Z)

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.