aiqbench analysis
This paper proposes Directed Temporal Representations for Control (DTRC), a method that refines latent geometry in frozen world models to better align with temporal reachability rather than just predictive similarity. By constructing a directed temporal quasimetric and using it as a critic, the approach achieves strong goal-conditioned control performance across ten visual tasks without requiring iterative trajectory search at test time. The work advances offline reinforcement learning by integrating temporal supervision directly into flow-policy parameterization to improve long-horizon planning capabilities.
Directed Temporal Representations for Offline Visual Control
arXiv:2610.08960v1 Announce Type: new Abstract: Predictive world models provide compact visual representations for control. Control requires a latent geometry aligned with temporal reachability rather than predictive similarity alone. We introduce Directed Temporal Representations for Control (DTRC), which learns such a geometry from offline visual trajectories on top of frozen LeWorldModel (LeWM) features. DTRC constructs a directed temporal quasimetric over the learned control representation. Short-range temporal offsets calibrate the distance scale. Bootstrapped targets extend temporal reachability across longer horizons. Action-conditioned consistency aligns the representation with local transition dynamics. The resulting distance estimates temporal reaching cost, and its change across a transition defines goal-relative temporal progress. We use this progress signal as a temporal critic for direct goal-conditioned policy learning. Model-assisted targets provide an additional training-time refinement under behavior-support and dynamics-agreement constraints. Across ten visual control tasks, DTRC achieves strong goal-conditioned control performance relative to planning and direct-policy baselines. Held-out diagnostics on the four LeWM tasks show consistent short-range temporal calibration, task-dependent long-range and directional structure, and positive transition-level progress. Temporal supervision improves the same flow-policy parameterization across all four LeWM tasks, while the resulting policy acts directly without iterative trajectory search at test time.
Abstract:Predictive world models provide compact visual representations for control. Control requires a latent geometry aligned with temporal reachability rather than predictive similarity alone. We introduce Directed Temporal Representations for Control (DTRC), which learns such a geometry from offline visual trajectories on top of frozen LeWorldModel (LeWM) features. DTRC constructs a directed temporal quasimetric over the learned control representation. Short-range temporal offsets calibrate the distance scale. Bootstrapped targets extend temporal reachability across longer horizons. Action-conditioned consistency aligns the representation with local transition dynamics. The resulting distance estimates temporal reaching cost, and its change across a transition defines goal-relative temporal progress. We use this progress signal as a temporal critic for direct goal-conditioned policy learning. Model-assisted targets provide an additional training-time refinement under behavior-support and dynamics-agreement constraints. Across ten visual control tasks, DTRC achieves strong goal-conditioned control performance relative to planning and direct-policy baselines. Held-out diagnostics on the four LeWM tasks show consistent short-range temporal calibration, task-dependent long-range and directional structure, and positive transition-level progress. Temporal supervision improves the same flow-policy parameterization across all four LeWM tasks, while the resulting policy acts directly without iterative trajectory search at test time.
Machine Learning (cs.LG)
arXiv:2610.08960 [cs.LG]
(or arXiv:2610.08960v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.08960 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 282 words
Related coverage
4DStreamCtrl: Interactive Video Generation with Online 4D Control
arXiv · Aug 28, 2026
Observation-Grounded Self-Predictive Reinforcement Learning for Visual Continuous Control
arXiv · Aug 7, 2026
Temporal-Distance JEPA: Plan-Aware Representation Learning for Latent World Model Predictive Control
arXiv · Jul 29, 2026
SyncRA: Learning Temporal Correspondence in Omni-Modal Models
arXiv · Sep 29, 2026