Dual Advantage Fields

Dual Advantage Fields

Alexey Zemtsov, Maxim Bobrin, Alexander Nikulin, Dmitry V. Dylov, Fakhri Karray, Vladislav Kurenkov, Martin Takáč, Arip Asadulaev · N/A · 2026

Framework

N/A

License

N/A

Stars

N/A

Summary

Offline goal-conditioned reinforcement learning requires both long-horizon reachability estimates and local action comparisons. Dual goal representations provide value fields that capture global goal reachability, but they do not directly specify which action should be preferred at a given state. We...

Abstract Summary

Offline goal-conditioned reinforcement learning requires both long-horizon reachability estimates and local action comparisons. Dual goal representations provide value fields that capture global goal reachability, but they do not directly specify which action should be preferred at a given state. We propose Dual Advantage Fields, a policy-extraction method that turns a bilinear dual value model into a local advantage signal. Under bilinear dual parameterization, the goal embedding is the gradient of the value field with respect to the state representation. DAF learns an action-effect model that predicts the discounted feature displacement induced by an action and scores actions by the alignment between this displacement and the goal direction. In the realizable case, this score equals the goal-conditioned Bellman advantage, yielding a standard local policy-improvement guarantee. On OGBench locomotion, manipulation, and puzzle tasks, DAF improves aggregate RLiable metrics and performs strongly in settings where locally correct actions differ from direct movement toward the final goal.

Key Points

  • Improves locomotion stability and agility
  • Addresses dexterous manipulation challenges
  • Incorporates tactile or force feedback for robust interaction
  • Learns multimodal motion skills from human demonstrations
  • Theoretically studies value/advantage decomposition for RL

Dual Advantage Fields

Authors: Alexey Zemtsov, Maxim Bobrin, Alexander Nikulin, Dmitry V. Dylov, Fakhri Karray, Vladislav Kurenkov, Martin Takáč, Arip Asadulaev

Venue: ICML 2026 | Year: 2026

arXiv: 2606.04188v1

Abstract

Offline goal-conditioned reinforcement learning requires both long-horizon reachability estimates and local action comparisons. Dual goal representations provide value fields that capture global goal reachability, but they do not directly specify which action should be preferred at a given state. We propose Dual Advantage Fields, a policy-extraction method that turns a bilinear dual value model into a local advantage signal. Under bilinear dual parameterization, the goal embedding is the gradient of the value field with respect to the state representation. DAF learns an action-effect model that predicts the discounted feature displacement induced by an action and scores actions by the alignment between this displacement and the goal direction. In the realizable case, this score equals the goal-conditioned Bellman advantage, yielding a standard local policy-improvement guarantee. On OGBench locomotion, manipulation, and puzzle tasks, DAF improves aggregate RLiable metrics and performs strongly in settings where locally correct actions differ from direct movement toward the final goal.

Key Contributions

  • Improves locomotion stability and agility
  • Addresses dexterous manipulation challenges
  • Incorporates tactile or force feedback for robust interaction
  • Learns multimodal motion skills from human demonstrations
  • Theoretically studies value/advantage decomposition for RL

Topics

  • reinforcement-learning
  • manipulation
  • locomotion

Code & Data

No code repository linked in paper metadata.

BibTeX

@article{Zemtsov2026_260604188v1,
  title={Dual Advantage Fields},
  author={Alexey Zemtsov and Maxim Bobrin and Alexander Nikulin and Dmitry V. Dylov and Fakhri Karray and Vladislav Kurenkov and Martin Takáč and Arip Asadulaev},
  year={2026},
  eprint={2606.04188v1},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  url={https://arxiv.org/abs/2606.04188v1}
}
Share

Related Papers

Accelerating and Scaling MPC-Guided Reinforcement Learning for Humanoid Locomotion and Manipulation

Accelerating and Scaling MPC-Guided Reinforcement Learning for Humanoid Locomotion and Manipulation

Junheng Li, Liang Wu, Sergio A. Esteban et al. · arXiv preprint · Jun 2026

In humanoid motion control, model predictive control (MPC) offers physically grounded prediction and constraint handling, while reinforcement learning (RL) enables robust whole-body skills through large-scale simulation. However, using MPC inside RL often requires time-consuming problem construction or excessive training overhead, making such frameworks difficult to justify in practice. This work studies efficient training-time MPC guidance for humanoid locomotion and manipulation, termed MPC-RL. We introduce a centroidal-dynamics MPC reward formulation that leverages guidance from MPC trajectories in training time. To make this practical in massively parallel RL, we develop π^nMPC, a parallel-in-horizon and construction-free batched GPU MPC solver that operates directly on time-varying dynamics to avoid high memory usage and pre-compilation. Through a variety of comparative studies and hardware validations, we have found that MPC-RL achieves superior performance in locomotion and manipulation skills.

humanoid reinforcement-learning model-predictive-control locomotion manipulation
Code PDF Advanced
GitHub ★ — Jun 2026
DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile Manipulation

DECOWAM: Decoupled Whole-Body World-Action Model for Legged Mobile Manipulation

Siyuan Ma, Boshi Zhang, Yutian Zhang et al. · arXiv preprint · Aug 2026

Mobile manipulation requires a robot to predict how locomotion and arm motion jointly alter future observations and control. Existing world-action models, developed largely for fixed-base platforms, do not explicitly distinguish camera ego-motion from base and arm actions. Here we introduce DECOW...

manipulation locomotion vision reinforcement-learning control benchmark
PDF Intermediate
No code repo Aug 2026
M3imic: Learning a Versatile Whole-Body Controller for Multimodal Motion Mimicking

M3imic: Learning a Versatile Whole-Body Controller for Multimodal Motion Mimicking

Zuxing Lu, Ziang Zheng, Yao Lyu et al. · arXiv preprint · Jun 2026

Building a general-purpose whole-body controller is essential for enabling diverse motion capabilities in humanoid robots across a wide range of downstream tasks, including locomotion and loco-manipulation. Different tasks rely on distinct motion reference modalities: locomotion primarily depends on...

sim-to-real reinforcement-learning locomotion manipulation humanoid
Code PDF Advanced
GitHub ★ — Jun 2026
Sample-Efficient Diffusion-based Reinforcement Learning with Critic Guidance

Sample-Efficient Diffusion-based Reinforcement Learning with Critic Guidance

Shutong Ding, Zejia Zhong, Zhongyi Wang et al. · ICML 2026 · May 2026

Recent advances in reinforcement learning (RL) have achieved great successes by leveraging the multimodality and exploration capability of diffusion policies. Among these approaches, one representativ

diffusion-policy reinforcement-learning locomotion manipulation
PDF Advanced
No code repo Code updated: May 2026