Related papers: Decision-Metric Alignment in Latent World Models: Diagnostics and Action-Conditioned Objectives for MPC Planning

Decision-Metric Alignment in Latent World Models: Diagnostics and Action-Conditioned Objectives for MPC Planning

URL: http://arxiv.org/abs/2608.18746v1
Date: Wed, 19 Aug 2026 09:56:39 GMT
Title: Decision-Metric Alignment in Latent World Models: Diagnostics and Action-Conditioned Objectives for MPC Planning
Abstract summary: JEPA-style latent world models can use Euclidean distance to a goal latent as the cost for model-predictive control (MPC)<n>We introduce Plan-Real Spearman, which measures latent--real rank agreement on random plans, and CEM-stage Spearman, which measures the same agreement as cross-entropy-method (CEM) search concentrates its proposal.
Score: 5.704305798718388
License: http://creativecommons.org/licenses/by/4.0/
Abstract: JEPA-style latent world models can use Euclidean distance to a goal latent as the cost for model-predictive control (MPC). Strong decoding of task variables, however, does not guarantee that this particular cost ranks candidate action sequences by real task progress. We call the latter property \emph{decision-metric alignment}. We introduce Plan-Real Spearman, which measures latent--real rank agreement on random plans, and CEM-stage Spearman, which measures the same agreement as cross-entropy-method (CEM) search concentrates its proposal. We analyze sufficient conditions under which latent distance preserves real-cost rankings, identifying encoder distortion, terminal rollout error, and candidate margins as the controlling quantities. Guided by the observed empirical alignment gap, DA-LeWM augments LeWM with inverse-dynamics and demonstration-conditioned goal-action heads. Across all our experiments, DA-LeWM accelerates convergence and achieves higher online success than LeWM, while probe scores remain similar. These results show that action-conditioned objectives improve the geometry used by Euclidean-cost, CEM-based latent MPC.

Related papers

Latent Energy Action Planning with World Models [16.141037661392115]
We introduce Latent Energy Action Planning (LEAP), which treats the complete action horizon as a differentiable variable.<n>LEAP couples terminal latent goal matching with a terminal-window state energy.<n>A quasi-Newton solver refines actions through the autoregressive rollout, and post-optimization projection enforces the admissible action range.
arXiv Detail & Related papers (2026-09-03T02:41:36Z)
SCALE: State-Calibrated Latent Embeddings for JEPA Planning in the Right Geometry [14.742804674497128]
Two strategies for obtaining non-collapsed representations are to inherit a pretrained feature space, and to learn an embedding end to end with anti-collapse regularization.<n>We show that planning depends not only on whether task-relevant information is present, but also on whether it shapes the geometry consumed by the planner.
arXiv Detail & Related papers (2026-08-17T08:58:55Z)
IADD-TR: Intervention-Aware Dynamics Decoupling with Targeted Regularization for Model-Based Reinforcement Learning [57.16513298507272]
We propose IADD-TR, a unified framework combining Intervention-Aware Dynamics Decoupling (IADD) and Targeted Regularization (TR)<n>IADD factorizes transitions into an action-intervention stage and an action-free natural evolution stage, using a zero-action anchor to resolve the non-uniqueness of this two-stage factorization for robust generalization.<n>For policy learning, we derive TR from the efficient influence function of a replay-state policy-gradient functional.
arXiv Detail & Related papers (2026-08-11T08:20:02Z)
Inference Time Causal Probing in LLMs [26.69306706280534]
In causal probing, an intervention modifies hidden states so that a property takes on a different value.<n>We propose Hidden-state Driven Margin Intervention (HDMI), a probe-free, gradient-based technique that directly steers hidden states.<n> HDMI consistently achieves higher reliability than prior methods on the LGD agreement corpus and the CausalGym benchmark.
arXiv Detail & Related papers (2026-05-08T11:59:13Z)
AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models [60.04879435087352]
Vision-language-action (VLA) policies generate actions within a single unified space.<n>We propose AnchorRefine, a hierarchical framework that factorizes VLA action modeling into trajectory anchor and residual refinement.<n> Experiments on LIBERO, CALVIN, and real-robot tasks demonstrate that AnchorRefine consistently improves both regression-based and diffusion-based VLA backbones.
arXiv Detail & Related papers (2026-04-20T04:25:24Z)
Grounded World Model for Semantically Generalizable Planning [94.53923128709965]
We learn a Grounded World Model (GWM) in a vision-language-aligned latent space.<n>Each proposed action is scored based on how close its future outcome is to the task instruction.<n>On the proposed WISER benchmark, GWM-MPC achieves a 87% success rate on the test set.
arXiv Detail & Related papers (2026-04-13T17:25:41Z)
Reshaping Action Error Distributions for Reliable Vision-Language-Action Models [69.38615670891038]
In robotic manipulation, vision-language-action (VLA) models have emerged as a promising paradigm for learning generalizable and scalable robot policies.<n>We focus on continuous-action VLA models and move beyond conventional MSE-based regression by reshaping action error distributions during training.<n>We evaluate our approaches across standard, few-shot, and noisy settings on multiple representative VLA architectures.
arXiv Detail & Related papers (2026-02-04T05:37:09Z)
MMDCP: A Distribution-free Approach to Outlier Detection and Classification with Coverage Guarantees and SCW-FDR Control [6.429952624399788]
We propose a unified framework for multi-class classification and outlier detection under label shift.<n>The Modified Mahalanobis Distance Conformal Prediction (MMDCP) combines class-specific distance measures with full conformal prediction to construct a score function.<n>We provide the first theoretical characterization of the gap between oracle and empirical conformal $p$-values, which ensures valid coverage and effective control of the class-wise false discovery rate (CW-FDR)
arXiv Detail & Related papers (2025-11-15T03:48:44Z)
Scalable Decision-Making in Stochastic Environments through Learned Temporal Abstraction [7.918703013303246]
We present Latent Macro Action Planner (L-MAP), which addresses the challenge of learning to make decisions in high-dimensional continuous action spaces.<n>L-MAP learns a set of temporally extended macro-actions through a state-conditional Vector Quantized Variational Autoencoder (VQ-VAE)<n>In offline RL settings, including continuous control tasks, L-MAP efficiently searches over discrete latent actions to yield high expected returns.
arXiv Detail & Related papers (2025-02-28T16:02:23Z)
When Demonstrations Meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement Learning [62.00672284480755]
This paper aims to recover the structure of rewards and environment dynamics that underlie observed actions in a fixed, finite set of demonstrations from an expert agent. Accurate models of expertise in executing a task has applications in safety-sensitive applications such as clinical decision making and autonomous driving.
arXiv Detail & Related papers (2023-02-15T04:14:20Z)
Modular Deep Reinforcement Learning for Continuous Motion Planning with Temporal Logic [59.94347858883343]
This paper investigates the motion planning of autonomous dynamical systems modeled by Markov decision processes (MDP) The novelty is to design an embedded product MDP (EP-MDP) between the LDGBA and the MDP. The proposed LDGBA-based reward shaping and discounting schemes for the model-free reinforcement learning (RL) only depend on the EP-MDP states.
arXiv Detail & Related papers (2021-02-24T01:11:25Z)

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