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

• URL: http://arxiv.org/abs/2607.26657v3
• Date: Thu, 06 Aug 2026 05:25:07 GMT
• Title: Enfold: Folding World Model Imagination into Predictive Representations for Ultra-Efficient Embodied Control
• Abstract summary: We argue that world generative models are most reusable when their computation that constructs a future is reusable.<n>We present Enfold, which transfers this computation into a representation predicted from the current visual context and language instruction.<n>Across LIBERO, RoboTwin2.0, and real-robot tasks, Enfold supports strong control while reducing action latency by $3.7times$ relative to Fast--WAM, Enfold-Flash reaches $10.1times$.
• Score: 41.15180918658398
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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.

Related papers

• GaussianDream++: Efficient 3D Gaussian World Modeling for Robotic Manipulation [55.47130289185285]
  Vision-Language-Action (VLA) policies have advanced language-conditioned robotic manipulation.<n>We present textbfMethodname, a compact, policy-native extension that inserts textbfWorld State Tokens and textbfWorld Prediction Tokens directly into the VLA backbone.
  arXiv  Detail & Related papers  (2026-08-26T11:41:55Z)
• ForgeDrive: Bidirectional Cross-Conditioning for Unified Visual-Action Generation in Autonomous Driving [12.68658021226119]
  We introduce ForgeDrive, a unified autoregressive diffusion framework with visual-action cross-conditioning.<n>We find that ForgeDrive not only unifies driving simulation, planning, and visual odometry into a single model, but also outperforms existing strong planners without any post-training strategy.
  arXiv  Detail & Related papers  (2026-06-30T07:05:48Z)
• MIRAGE: Mobile Agents with Implicit Reasoning and Generative World Models [53.09772037247959]
  We introduce MIRAGE, a framework that learns continuous latent reasoning representations from visible textual reasoning traces.<n>On AndroidWorld, MIRAGE matches explicit chain-of-thought supervised fine-tuning in the 4B ablation with a 3-5x lower decoded-token budget.<n>On AndroidControl, it improves action grounding while generating over 75% fewer tokens.
  arXiv  Detail & Related papers  (2026-06-03T09:01:24Z)
• From Imagined Futures to Executable Actions: Mixture of Latent Actions for Robot Manipulation [88.39072412680633]
  We propose MoLA, a control-oriented interface that transforms imagined future videos into executable representations.<n>We evaluate our approach on simulated benchmarks and real-world robot manipulation tasks.
  arXiv  Detail & Related papers  (2026-05-12T14:15:16Z)
• A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens [40.71386517238762]
  We introduce DeltaTok, a tokenizer that encodes the feature difference between consecutive frames into a single "delta" token.<n>We also present DeltaWorld, a generative world model operating on these tokens to efficiently generate diverse plausible futures.
  arXiv  Detail & Related papers  (2026-04-06T17:55:05Z)
• Parallel Decoding via Hidden Transfer for Lossless Large Language Model Acceleration [54.897493351694195]
  We propose a novel parallel decoding approach, namely textithidden transfer, which decodes multiple successive tokens simultaneously in a single forward pass. In terms of acceleration metrics, we outperform all the single-model acceleration techniques, including Medusa and Self-Speculative decoding.
  arXiv  Detail & Related papers  (2024-04-18T09:17:06Z)
• iTransformer: Inverted Transformers Are Effective for Time Series Forecasting [62.40166958002558]
  We propose iTransformer, which simply applies the attention and feed-forward network on the inverted dimensions. The iTransformer model achieves state-of-the-art on challenging real-world datasets.
  arXiv  Detail & Related papers  (2023-10-10T13:44:09Z)
• Future Sight: Dynamic Story Generation with Large Pretrained Language Models [11.23192733149335]
  Transformer decoders can only generate new text with respect to previously generated text. Future Sight enables a decoder to attend to an encoded future plot event. During inference, the future plot event can be written by a human author to steer the narrative being generated in a certain direction.
  arXiv  Detail & Related papers  (2022-12-20T01:53:26Z)
• Back to the Future: Unsupervised Backprop-based Decoding for Counterfactual and Abductive Commonsense Reasoning [79.48769764508006]
  generative language models (LMs) can be trained to condition only on the past context or to perform narrowly scoped text-infilling. We propose DeLorean, a new unsupervised decoding algorithm that can flexibly incorporate both the past and future contexts. We demonstrate that our approach is general and applicable to two nonmonotonic reasoning tasks: abductive text generation and counterfactual story revision.
  arXiv  Detail & Related papers  (2020-10-12T17:58:43Z)
• Addressing Some Limitations of Transformers with Feedback Memory [51.94640029417114]
  Transformers have been successfully applied to sequential, auto-regressive tasks despite being feedforward networks. We propose the Feedback Transformer architecture that exposes all previous representations to all future representations. We demonstrate on a variety of benchmarks in language modeling, machine translation, and reinforcement learning that the increased representation capacity can create small, shallow models with much stronger performance than comparable Transformers.
  arXiv  Detail & Related papers  (2020-02-21T16:37:57Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.