Back to Parsimonious Latents: Learning Task-Centric World Models from Visual Foundations

• URL: http://arxiv.org/abs/2605.25620v1
• Date: Mon, 25 May 2026 09:21:43 GMT
• Title: Back to Parsimonious Latents: Learning Task-Centric World Models from Visual Foundations
• Abstract summary: We propose a framework for turning foundation-model embeddings into compact, task-sufficient world representations.<n> TC-WM linearly projects high-dimensional visual embeddings into a compact latent as the dynamic space.<n>We show that TC-WM suffices to identify the underlying task-centric latent factors up to a simple transformation.
• Score: 35.521396983843864
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: World models enable agents to predict future dynamics conditioned on actions, making the choice of latent representation central to planning and control. Such representations are often either learned directly from pixels with limited semantic structure or inherited from frozen visual foundation models with excessive task-irrelevant detail, yielding state spaces that are poorly matched to downstream planning and control. This is especially challenging in reward-free offline settings, where the model must learn from fixed trajectories without reward supervision or online interaction. To address this, we propose TC-WM, a framework for turning foundation-model embeddings into compact, task-sufficient world representations. The key design is to treat the pretrained embedding space as a semantic scaffold rather than as the final state space: TC-WM linearly projects high-dimensional visual embeddings into a compact latent as the dynamic space, aligns a subspace with the agent's physical state via contrastive learning, and reconstructs embeddings to preserve useful visual structure. This combines the generality of foundation features with the controllability of task-centric dynamics. Theoretically, we show that TC-WM suffices to identify the underlying task-centric latent factors up to a simple transformation. Empirically, TC-WM enables test-time planning across diverse environments (e.g., Robomimic and D4RL), achieving better world-modeling quality and more precise control than state-of-the-art approaches.

Related papers

• Sensorimotor World Models: Perception for Action via Inverse Dynamics [59.121736411156384]
  We introduce a sensorimotor world model trained end-to-end with inverse dynamics regularization.<n>It prevents representation collapse and induces action-aligned representations.<n>It learns compact, interpretable latent spaces and enables competitive planning performance across simple 2D and 3D control tasks.
  arXiv  Detail & Related papers  (2026-06-18T11:25:16Z)
• BRICKS-WM: Building Reusability via Interface Composition Kinetics for Structured World Models [44.13098312901522]
  We introduce BRICKS-WM, a framework for the modular assembly of structured world models.<n>As a minimal instantiation, we factorize the latent state space into an actuated Agent module and an external Background module.<n>BRICKS-WM achieves control performance comparable to strong monolithic baselines when trained from scratch.
  arXiv  Detail & Related papers  (2026-06-15T09:55:33Z)
• The Art of Interrogation: Consistency Amplifies Factuality in Spatial Reasoning [65.67266333751569]
  Current Large Reasoning Models (LRMs) exhibit remarkable general capabilities but significantly underperform in spatial reasoning tasks.<n>We propose a self-supervised reinforcement learning framework that targets the internal reasoning process without requiring ground-truth annotations.<n>We show that this label-free consistency training approaches the accuracy of models trained with ground-truth supervision and achieves similar generalization across diverse tasks and data domains.
  arXiv  Detail & Related papers  (2026-06-10T10:50:06Z)
• AffordanceVLA: A Vision-Language-Action Model Empowering Action Generation through Affordance-Aware Understanding [55.535374902204296]
  AffordanceVLA is a unified framework that introduces structured affordance forecasting as a task-oriented intermediate representation.<n>We show that AffordanceVLA achieves strong performance across diverse manipulation scenarios.
  arXiv  Detail & Related papers  (2026-06-04T13:28:51Z)
• Telecom World Models: Unifying Digital Twins, Foundation Models, and Predictive Planning for 6G [60.29349053952935]
  We introduce the Telecom World Model(TWM) concept, an architecture for learned, action-conditioned, uncertainty-aware modeling of telecom system dynamics.<n>We propose a three-layer architecture, comprising a field world model for spatial environment prediction, a control/dynamics world model for action-conditioned Key Performance Indicator (KPI) trajectory prediction, and a telecom foundation model layer for intent translation and orchestration.
  arXiv  Detail & Related papers  (2026-04-08T09:41:58Z)
• Revisiting Multi-Task Visual Representation Learning [52.93947931352643]
  We introduce MTV, a principled multi-task visual pretraining framework.<n>We leverage high-capacity "expert" models to synthesize dense, structured pseudo-labels at scale.<n>Our results demonstrate that MTV achieves "best-of-both-worlds" performance.
  arXiv  Detail & Related papers  (2026-01-20T11:59:19Z)
• Prismatic World Model: Learning Compositional Dynamics for Planning in Hybrid Systems [38.4555621948915]
  Prismatic World Model (PRISM-WM) is designed to decompose complex hybrid dynamics into composable primitives.<n>PRISM-WM significantly reduces rollout drift by accurately modeling sharp mode transitions in system dynamics.
  arXiv  Detail & Related papers  (2025-12-09T09:40:34Z)
• GrndCtrl: Grounding World Models via Self-Supervised Reward Alignment [16.343768407636322]
  We introduce Reinforcement Learning with World Grounding (RLWG), a self-supervised post-training framework that aligns pretrained world models with a physically verifiable structure through geometric and perceptual rewards.<n>We instantiate this framework with GrndCtrl, a reward-aligned adaptation method based on Group Relative Policy Optimization (GRPO), yielding world models that maintain stable trajectories, consistent geometry, and reliable rollouts for embodied navigation.
  arXiv  Detail & Related papers  (2025-12-01T18:03:29Z)
• Unified World Models: Memory-Augmented Planning and Foresight for Visual Navigation [69.94565127141483]
  Current approaches separate navigation planning from visual world modeling, leading to state-action misalignment and limited adaptability.<n>We propose UniWM, a unified, memory-augmented world model integrating egocentric visual foresight and planning within a single multimodal autoregressive backbone.<n>We show that UniWM substantially improves navigation success rates by up to 30%, significantly reduces trajectory errors compared to strong baselines, and exhibits impressive zero-shot generalization on the unseen TartanDrive dataset.
  arXiv  Detail & Related papers  (2025-10-09T18:18:11Z)
• From Local Cues to Global Percepts: Emergent Gestalt Organization in Self-Supervised Vision Models [7.7536110932446265]
  We investigate whether modern vision models show similar behaviors, and under what training conditions these emerge.<n>We find that Vision Transformers (ViTs) trained with Masked Autoencoding (MAE) exhibit activation patterns consistent with Gestalt laws.<n>We introduce the Distorted Spatial Relationship Testbench (DiSRT), which evaluates sensitivity to global spatial perturbations while preserving local textures.
  arXiv  Detail & Related papers  (2025-05-31T21:35:54Z)
• Spatiotemporal Implicit Neural Representation as a Generalized Traffic Data Learner [46.866240648471894]
  Spatiotemporal Traffic Data (STTD) measures the complex dynamical behaviors of the multiscale transportation system. We present a novel paradigm to address the STTD learning problem by parameterizing STTD as an implicit neural representation. We validate its effectiveness through extensive experiments in real-world scenarios, showcasing applications from corridor to network scales.
  arXiv  Detail & Related papers  (2024-05-06T06:23:06Z)
• Continual Visual Reinforcement Learning with A Life-Long World Model [55.05017177980985]
  We present a new continual learning approach for visual dynamics modeling.<n>We first introduce the life-long world model, which learns task-specific latent dynamics.<n>Then, we address the value estimation challenge for previous tasks with the exploratory-conservative behavior learning approach.
  arXiv  Detail & Related papers  (2023-03-12T05:08:03Z)
• Temporal Predictive Coding For Model-Based Planning In Latent Space [80.99554006174093]
  We present an information-theoretic approach that employs temporal predictive coding to encode elements in the environment that can be predicted across time. We evaluate our model on a challenging modification of standard DMControl tasks where the background is replaced with natural videos that contain complex but irrelevant information to the planning task.
  arXiv  Detail & Related papers  (2021-06-14T04:31:15Z)

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.