paper-with-me

Papers

Drifting Field Policy: A One-Step Generative Policy via Wasserstein Gradient Flow

2026-05-08 · Juil Koo, Mingue Park, Jiwon Choi, Yunhong Min, Minhyuk Sung arxiv

We propose Drifting Field Policy (DFP), a non-ODE one-step generative policy built on the drifting model paradigm. We frame the policy update as a reverse-KL Wasserstein-2 gradient flow toward a soft target policy, so that each DFP update corresponds to a gradient step in probability space. By construction, this gradient is decomposed into an ascent toward higher action-value regions and a score matching with the anchor policy as a trust region. We further derive a simple, tractable surrogate of the otherwise intractable update loss, akin to behavior cloning on top-K critic-selected actions. We find empirically that this mechanism uniquely benefits the drifting backbone owing to its non-ODE parameterization. With one-step inference, DFP achieves state-of-the-art performance on several manipulation tasks across Robomimic and OGBench, outperforming ODE-based policies.

📄 PDF Abstract BibTeX arXiv:2605.07727

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Implicit Drifting Policy: One-Step Action Generation via Conditional Expert Geometry

2026-05-31 · Zemin Yang, Yaoyu He, Yiming Zhong, Yuhao Zhang 외 arxiv

Generative action policies based on diffusion or flow matching excel in behavior cloning, yet their iterative sampling is prohibitive for high-frequency robot control. While recent one-step formulations alleviate this la…

Positive-Only Drifting Policy Optimization

2026-04-15 · Qi Zhang arxiv

In the field of online reinforcement learning (RL), traditional Gaussian policies and flow-based methods are often constrained by their unimodal expressiveness, complex gradient clipping, or stringent trust-region requir…

Reinforcement Learning

Drift-Based Policy Optimization: Native One-Step Policy Learning for Online Robot Control

2026-04-04 · Yuxuan Gao, Yedong Shen, Shiqi Zhang, Wenhao Yu 외 arxiv

Although multi-step generative policies achieve strong performance in robotic manipulation by modeling multimodal action distributions, they require multi-step iterative denoising at inference time. Each action therefore…

Reinforcement Learning

DriftXpress: Faster Drifting Models via Projected RKHS Fields

2026-05-12 · Ali Falahati, Elliot Creager, Gautam Kamath, Shubhankar Mohapatra arxiv

Drifting Models have emerged as a new paradigm for one-step generative modeling, achieving strong image quality without iterative inference. The premise is to replace the iterative denoising process in diffusion models w…

Aligning Few-Step Generative Models by Amortizing Sample-based Variational Inference

2026-05-26 · Jaewoo Lee, Hyeongyu Kang, Dohyun Kim, Kyuil Sim 외 arxiv

Aligning a few-step generative model is challenging, since existing alignment frameworks typically rely on restrictive assumptions: a tractable likelihood, a specific ODE/SDE solver, or a particular model family. We intr…