What Matters for Latent Actions in Robot Learning

• URL: http://arxiv.org/abs/2608.19613v1
• Date: Thu, 20 Aug 2026 03:54:51 GMT
• Title: What Matters for Latent Actions in Robot Learning
• Abstract summary: Latent Action Models (LAMs) have emerged as a promising paradigm for enabling robot learning to leverage large-scale unlabeled videos.<n>We present the first comprehensive empirical study of latent action learning for robotic manipulation.
• Score: 41.733127985154034
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Latent Action Models (LAMs) have emerged as a promising paradigm for enabling robot learning to leverage large-scale unlabeled videos through latent actions that serve as compact surrogates for physical actions. Despite rapid progress, research on LAM remains highly fragmented, with existing methods evaluating different design choices in isolation under inconsistent experimental settings, making it difficult to identify the factors that truly determine downstream robotic manipulation performance. In this work, we present the first comprehensive empirical study of latent action learning for robotic manipulation. We unify representative LAM methods within a common autoencoding framework and systematically investigate 41 LAM design choices across three dimensions, including latent action modeling paradigms, learning objectives and regularization methods, and latent action integration strategies. We further examine four proxy metrics for evaluating latent action quality and assess their ability to reliably predict downstream robotic manipulation performance. Extensive experiments on three widely used benchmarks provide strong empirical evidence that fine-tuning vision-language model (VLM) backbones with latent actions provides a stronger initialization for downstream policy learning, with further validation on real-world robot manipulation tasks.

Related papers

• Simple-to-Complex Structured Demonstrations for Vision-Language-Action Learning [9.340058165047717]
  Vision-Language-Action (VLA) models have demonstrated strong capabilities in robotic manipulation.<n>We identify demonstration organization as a fundamental yet overlooked aspect of imitation learning.<n>We propose a simple-to-complex structured demonstration collection strategy for VLA learning using a dual-arm robotic platform.
  arXiv  Detail & Related papers  (2026-07-06T01:38:43Z)
• Reshaping Action Error Distributions for Reliable Vision-Language-Action Models [69.38615670891038]
  In robotic manipulation, vision-language-action (VLA) models have emerged as a promising paradigm for learning generalizable and scalable robot policies.<n>We focus on continuous-action VLA models and move beyond conventional MSE-based regression by reshaping action error distributions during training.<n>We evaluate our approaches across standard, few-shot, and noisy settings on multiple representative VLA architectures.
  arXiv  Detail & Related papers  (2026-02-04T05:37:09Z)
• Embodied Robot Manipulation in the Era of Foundation Models: Planning and Learning Perspectives [101.22313265304979]
  Recent advances in vision, language, and multimodal learning have accelerated progress in robotic foundation models.<n>This survey examines robot manipulation from an algorithmic perspective.<n>We organize recent learning-based approaches within a unified abstraction of high-level planning and low-level control.
  arXiv  Detail & Related papers  (2025-12-28T16:05:38Z)
• StaMo: Unsupervised Learning of Generalizable Robot Motion from Compact State Representation [56.996371714721995]
  We propose an unsupervised approach that learns a highly compressed two-token state representation.<n>Our representation is efficient, interpretable, and integrates seamlessly into existing VLA-based models.<n>We name our method StaMo for its ability to learn generalizable robotic Motion from compact State representation.
  arXiv  Detail & Related papers  (2025-10-06T17:37:24Z)
• 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)
• Evaluating Uncertainty and Quality of Visual Language Action-enabled Robots [13.26825865228582]
  We propose eight uncertainty metrics and five quality metrics specifically designed for VLA models for robotic manipulation tasks.<n>We assess their effectiveness through a large-scale empirical study involving 908 successful task executions from three state-of-the-art VLA models.
  arXiv  Detail & Related papers  (2025-07-22T22:15:59Z)
• Active-O3: Empowering Multimodal Large Language Models with Active Perception via GRPO [63.140883026848286]
  Active vision refers to the process of actively selecting where and how to look in order to gather task-relevant information.<n>Recently, the use of Multimodal Large Language Models (MLLMs) as central planning and decision-making modules in robotic systems has gained extensive attention.
  arXiv  Detail & Related papers  (2025-05-27T17:29:31Z)
• Sample Efficient Robot Learning in Supervised Effect Prediction Tasks [0.0]
  MUSEL (Model Uncertainty for Sample-Efficient Learning) is a novel AL framework tailored for regression tasks in robotics.<n>We show that MUSEL improves both learning accuracy and sample efficiency, validating its effectiveness in learning action effects selecting informative samples.
  arXiv  Detail & Related papers  (2024-12-03T09:48:28Z)
• Tactile Active Inference Reinforcement Learning for Efficient Robotic Manipulation Skill Acquisition [10.072992621244042]
  We propose a novel method for skill learning in robotic manipulation called Tactile Active Inference Reinforcement Learning (Tactile-AIRL) To enhance the performance of reinforcement learning (RL), we introduce active inference, which integrates model-based techniques and intrinsic curiosity into the RL process. We demonstrate that our method achieves significantly high training efficiency in non-prehensile objects pushing tasks.
  arXiv  Detail & Related papers  (2023-11-19T10:19:22Z)
• Sample-Efficient Reinforcement Learning with Symmetry-Guided Demonstrations for Robotic Manipulation [7.099237102357281]
  Reinforcement learning (RL) suffers from low sample efficiency, particularly in high-dimensional continuous state-action spaces.<n>We introduce Demo-EASE, a novel training framework using a dual-buffer architecture that stores both demonstrations and RL-generated experiences.<n>Our results show that Demo-EASE significantly accelerates convergence and improves final performance compared to standard RL baselines.
  arXiv  Detail & Related papers  (2023-04-12T11:38:01Z)
• Active Exploration for Robotic Manipulation [40.39182660794481]
  This paper proposes a model-based active exploration approach that enables efficient learning in sparse-reward robotic manipulation tasks. We evaluate our proposed algorithm in simulation and on a real robot, trained from scratch with our method.
  arXiv  Detail & Related papers  (2022-10-23T18:07:51Z)

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.