Efficient Inference for World Action Models

Follow this topic Ask about this topic

World action models predict actions together with how the world may change, and efficient inference studies how to run these models within real-time and memory limits. Approaches include using compact latent, motion, or object representations instead of dense video, distilling models to generate actions in few steps, and reusing cached computations across control cycles. Systems also use pruning, quantization, asynchronous streaming, and adaptive compute, with evaluation on robot manipulation benchmarks such as LIBERO and RoboTwin as well as real robots and navigation tasks. The central tension is that future prediction can aid robustness and long-horizon planning, but generating or reusing it adds cost and can introduce stale or inaccurate information.

Latest papers

80 papers37 papers in the last four weeks, up 164% on the four weeks before. 0.2% of all new papers.

Enfold: Folding World Model Imagination into Predictive Representations for Ultra-Efficient Embodied Control
Jul 29, 2026Weili Zeng, Yitong Xing, Fulong Liu +10World Models Efficient VLA Model Inference
Shanghai Jiao Tong University · Qilu University of Technology (Shandong Academy of Sciences)
ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU
Jul 21, 2026Fan Jiang, Zhaoxu Sun, Mengchao Wang +38Action-Conditioned Video Generation Long-Horizon Video Generation
DriftWorld: Fast World Modeling through Drifting
Jul 16, 2026Susie Lu, Haonan Chen, Weirui Ye +1World Models for Robotics World Model-Based Planning
Massachusetts Institute of Technology · Harvard University
GigaWorld-Policy-0.5: A Faster and Stronger WAM Empowered by AutoResearch
Jul 15, 2026GigaWorld Team, Angen Ye, Angyuan Ma +26Robot Policy Learning World Model Learning
GigaAI Tsinghua University
Fast LeWorldModel
Jun 24, 2026Yuntian Gao, Xiangyu XuLatent Dynamics Modeling World Model-Based Planning
Xi’an Jiaotong University
MemoryWAM: Efficient World Action Modeling with Persistent Memory
Jun 18, 2026Sizhe Yang, Juncheng Mu, Tianming Wei +8Long-Horizon Robotic Manipulation World Models for Robotics
The Chinese University of Hong Kong · Tsinghua University · Zhejiang University
Efficient-WAM: A 1B-Parameter World-Action Model with Low-Cost Future Imagination
Jun 8, 2026Jiajun Li, Tiecheng Guo, Yifan Ye +9World Models for Robotics Efficient Inference for World Action Models
1The University of Hong Kong · 2Peking University · 3Muka Robotics +2
C $^3$ ache: Accelerating World Action Models with Cross Inference Chunk Cache
Jun 8, 2026Weisen Zhao, Lam Nguyen, Zhicong Lu +1World Model Learning Action-Conditioned World Models
George Mason University · University of Central Florida
Light-WAM: Efficient World Action Models with State-Fusion Action Decoding
Jun 6, 2026Ziang Li, Dongzhou Cheng, Yibin Wang +5Robot Policy Learning World Models for Robotics
Wuhan University · Shanghai Innovation Institute · Southeast University +2
Flash-WAM: Modality-Aware Distillation for World Action Models
Jun 3, 2026Arman Akbari, Ci Zhang, Arash Akbari +6Video Diffusion Models Action-Conditioned World Models
1Northeastern University · University of Georgia · 3EmbodyX Inc.
WorldKV: Efficient World Memory with World Retrieval and Compression
May 21, 2026Jung Yi, Minjae Kim, Paul Hyunbin Cho +3KV Caching Memory-Augmented Video Generation
1KAIST AI · 2Naver AI Lab
SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformer
May 14, 2026Haoyi Zhu, Haozhe Liu, Yuyang Zhao +6Camera-Controlled Video Generation Video Diffusion Models
NVIDIA
Motubrain: An Advanced World Action Model for Robot Control
Apr 30, 2026Motubrain Team, Chendong Xiang, Fan Bao +17Robotic Control World Models for Robotics
Motubrain Team
DexWorldModel: Causal Latent World Modeling towards Automated Learning of Embodied Tasks
Apr 13, 2026Yueci Deng, Guiliang Liu, Kui JiaWorld Models for Robotics Latent World Models
DexForce AI
Describe-Then-Act: Proactive Agent Steering via Distilled Language-Action World Models
Mar 24, 2026Massimiliano Pappa, Luca Romani, Valentino Sacco +5Efficient Inference for World Action Models AI Agent Safety
Sapienza University of Rome, Italy · Inria, Univ. Grenoble Alpes, CNRS, LJK
GigaWorld-Policy: An Efficient Action-Centered World--Action Model
Mar 18, 2026Angen Ye, Boyuan Wang, Chaojun Ni +21World Action Models Visuomotor Policy Learning
Fast-WAM: Do World Action Models Need Test-time Future Imagination?
Mar 17, 2026Tianyuan Yuan, Zibin Dong, Yicheng Liu +1World Action Models Efficient Inference for World Action Models
IIIS, Tsinghua University · Galaxea AI
MWM: Mobile World Models for Action-Conditioned Consistent Prediction
Mar 8, 2026Han Yan, Zishang Xiang, Zeyu Zhang +1World Models for Robotics Action-Conditioned World Models
School of Computer Science, Peking University
Planning in 8 Tokens: A Compact Discrete Tokenizer for Latent World Model
Mar 5, 2026Dongwon Kim, Gawon Seo, Jinsung Lee +2World Model-Based Planning Efficient Inference for World Action Models
KAIST · POSTECH · RLWRLD
World Action Models are Zero-shot Policies
Feb 17, 2026Seonghyeon Ye, Yunhao Ge, Kaiyuan Zheng +33World Action Models Zero-Shot Robotic Manipulation
NVIDIA

Jul 29, 2026J/K move · Enter open · S save

Enfold: Folding World Model Imagination into Predictive Representations for Ultra-Efficient Embodied Control

Weili Zeng, Yitong Xing, Fulong Liu, Chengqun Yang+9

Shanghai Jiao Tong University · Qilu University of Technology (Shandong Academy of Sciences)

World Models Efficient VLA Model Inference Efficient Inference for World Action Models Vision-Language-Action Models Predictive Representation Learning Embodied World Models

Open paper AskPDF arXiv

World generative models are typically used through what they produce: a rendered future, a video-conditioned action, or latent context computed by a costly generative branch. We argue that their more reusable asset is the computation that constructs a future. As a generator transforms a corrupted future into a coherent trajectory, its intermediate states organize appearance, spatial layout, and interaction across levels of abstraction. Can this future-generative computation be internalized in a representation inferred from the present alone? We present Enfold, which transfers this computation into a representation predicted from the current visual context and language instruction. During training, multi-level states exposed as the generator processes the observed future supervise a current-only encoder. The learned representation is fed back to condition future generation and is read by task heads without allowing task gradients to reshape the encoder. At deployment, action prediction no longer executes the generator. Across LIBERO, RoboTwin2.0, and real-robot tasks, Enfold supports strong control while reducing action latency by

3.7\times

relative to Fast--WAM, Enfold-Flash reaches

10.1\times

. Representation analyses show that it suppresses nuisance variation and preferentially captures changes that emerge over longer horizons. When the current scene is altered by human intervention, both the generated continuation and the executed actions adapt, which is inconsistent with fixed trajectory replay. These results recast a world generator as a source of predictive control representations: its future need not be materialized at every step if its internal structure can be enfolded into the present.