publications

2026

1. Preprint

   Trust Region Q Adjoint Matching

   Yonghoon Dong, Kyungmin Lee, Changyeon Kim, Jaehyuk Kim, and Jinwoo Shin

   arXiv preprint arXiv:2605.27079, 2026

   Abs arXiv Bib Code Website

   Off-policy reinforcement learning of pretrained flow policies remains challenging due to the instability of optimization arising from the multi-step sampling process. Recently, Q-learning with Adjoint Matching (QAM) addressed this issue by reformulating into a memoryless stochastic optimal control (SOC) problem with a learned critic. However, QAM inherits a fundamental fragility of critic-guided improvement: small critic errors are amplified when critics are ill-conditioned, often leading to model collapse. This paper introduces Trust Region Q-Adjoint Matching (TRQAM), a stable off-policy fine-tuning algorithm that adaptively controls the path-space KL with pretrained flow policies through projected dual descent. Specifically, we optimize the trust-region parameter λin SOC dynamics, and theoretically show that the path-space KL can be represented by a closed-form function of λ. As a result, our method can precisely control the exact deviation from pretrained flow policies, achieving stable off-policy RL. Through experiments on 50 OGBench tasks, TRQAM consistently outperforms prior arts in both offline RL and offline-to-online RL. In particular, TRQAM achieves an overall success rate of 68% in offline RL, substantially improves the strongest baseline at 46%.

   @article{dong2026trqam,
     title = {Trust Region Q Adjoint Matching},
     author = {Dong, Yonghoon and Lee, Kyungmin and Kim, Changyeon and Kim, Jaehyuk and Shin, Jinwoo},
     journal = {arXiv preprint arXiv:2605.27079},
     year = {2026},
   }

2. Tech Report

   RLDX-1 Technical Report

   Dongyoung Kim^*, Huiwon Jang^*, Myungkyu Koo^†, Suhyeok Jang^†, Taeyoung Kim^†, Yonghoon Dong, and Jinwoo Shin

   arXiv preprint arXiv:2605.03269, 2026

   Abs arXiv Bib Code Website

   (* Project Lead, † Research Lead.) While Vision-Language-Action models (VLAs) have shown remarkable progress toward human-like generalist robotic policies through the versatile intelligence (i.e., broad scene understanding and language-conditioned generalization) inherited from pre-trained Vision-Language Models, they still struggle with complex real-world tasks requiring broader functional capabilities (e.g., motion awareness, long-term memory, and physical sensing). To address this, we introduce RLDX-1, a general-purpose robotic policy for dexterous manipulation built on the Multi-Stream Action Transformer (MSAT), an architecture that unifies these capabilities by integrating heterogeneous modalities through modality-specific streams with cross-modal joint self-attention. RLDX-1 further combines this architecture with system-level design choices, including data synthesis for rare manipulation scenarios, learning procedures specialized for human-like manipulation, and inference optimizations for real-time deployment. Through empirical evaluation, we show that RLDX-1 consistently outperforms recent frontier VLAs (e.g., \pi_0.5 and GR00T N1.6) across both simulation benchmarks and real-world tasks that require broad functional capabilities beyond general versatility. In particular, RLDX-1 shows superiority in ALLEX humanoid tasks by achieving success rates of 86.8% while \pi_0.5 and GR00T N1.6 achieve around 40%, highlighting the ability of RLDX-1 to control a high-DoF humanoid robot under diverse functional demands. Together, these results position RLDX-1 as a promising step toward reliable VLAs for complex, contact-rich, and dynamic real-world dexterous manipulation.

   @article{kim2026rldx1,
     title = {RLDX-1 Technical Report},
     author = {Kim, Dongyoung and Jang, Huiwon and Koo, Myungkyu and Jang, Suhyeok and Kim, Taeyoung and Dong, Yonghoon and Shin, Jinwoo},
     journal = {arXiv preprint arXiv:2605.03269},
     year = {2026},
   }