paper-with-me

홈 › Papers

RISE: Adaptive Imagination for World Action Models

2026-08-20 · Hongbo Lu, Liang Yao, Chenghao He, Hao Han, Fan Liu, Wenlong Liao, Tao He, Pai Peng hf

World Action Models (WAMs) improve planning by incorporating future world evolution into action generation, yet existing methods allocate a fixed imagination budget to every scene. We propose RISE (Refining Imagination through SElective Rollout), a system-level adaptive imagination framework that makes sequential Roll/Stop decisions according to the expected planning benefit of continued rollout. At each step, a Latent Evaluator estimates the risk revealed by the current prefix and how much planning could improve if imagination continues, while a Rollout Gate weighs this expected benefit against additional computation cost. Since factual driving logs expose only one realized future, we further construct CounterDrive, a counterfactual dataset with diverse outcomes and risk levels, to enrich future dynamics and provide localized risk supervision. Each retained sample undergoes expert verification and annotation of trajectory validity, incident onset, and causal category, providing a reusable resource for safety-critical world-modeling research. Experiments on NAVSIM and nuScenes show that RISE achieves the best overall planning performance while reducing unnecessary rollout, with additional transfer results supporting its plug-in generality across WAM architectures.

📄 PDF Abstract BibTeX arXiv:2608.20430

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

When and How Much to Imagine: Adaptive Test-Time Scaling with World Models for Visual Spatial Reasoning

2026-02-09 · Shoubin Yu, Yue Zhang, Zun Wang, Jaehong Yoon 외 arxiv

Despite rapid progress in MLLMs, visual spatial reasoning remains unreliable when correct answers depend on how a scene would appear under unseen or alternative viewpoints. Recent work addresses this by augmenting reason…

Spatial Reasoning

Attacking the Trusted Imagination: Oracle-Level Integrity Attacks on Imagine-then-Act World Models

2026-06-22 · Linghan Chen, Kaiyan Ji, Minyu Guo arxiv

Many recent vision-language-action (VLA) policies adopt an imagine-then-act design. A world-action model (WAM) first imagines a short future as a latent trajectory z~, on which the action is then conditioned. We identify…

When to Trust Imagination: Adaptive Action Execution for World Action Models

2026-05-07 · Rui Wang, Yue Zhang, Jiehong Lin, Kuncheng Luo 외 arxiv

World Action Models (WAMs) have recently emerged as a promising paradigm for robotic manipulation by jointly predicting future visual observations and future actions. However, current WAMs typically execute a fixed numbe…

Domain Adaptation with Adaptive Imagination for Visual Reinforcement Learning under Limited Target Data

2026-06-29 · Hyunwoo Park, Sang-Hyun Lee arxiv

Sim-to-real transfer remains a major obstacle for reinforcement learning (RL), especially for vision-based control where image observations exacerbate the state-distribution shift between simulation and the real world. D…

Reinforcement LearningDomain Adaptation

Adaptive World Models: Learning Behaviors by Latent Imagination Under Non-Stationarity

2024-11-02 · Emiliyan Gospodinov, Vaisakh Shaj, Philipp Becker, Stefan Geyer 외

Developing foundational world models is a key research direction for embodied intelligence, with the ability to adapt to non-stationary environments being a crucial criterion. In this work, we introduce a new formalism, …