AI & ML
Beyond Prediction Gain: LASR as a Metric for Action-Grounded Latents in World Models
arXiv cs.RO0 views
arXiv:2611.17775v1 Announce Type: new Abstract: World models that support planning, control, and interpretable reasoning require latent representations that genuinely encode action. Yet the large-scale video data from which such models are learned overwhelmingly lack action annotations. Latent-action models infer action-like variables from observation sequences, enabling action-aware learning from unlabeled videos or trajectories. A fundamental concern, however, remains: a latent that improves future prediction may simply encode the current state, episode progress, task identity, or other predictive shortcuts. Existing evaluation practices---reconstruction loss, rollout fidelity, action recovery, or downstream performance---cannot isolate this distinction. We introduce the Latent Action Support Ratio (LASR), a dimensionless metric that measures, under offline evaluation on held-out data, whether a learned latent provides predictive support for action-related targets beyond the current state and non-action alternatives. LASR provides two complementary summaries: GeoLASR captures typical support, while WorstLASR identifies the weakest case. We instantiate the LASR benchmark with recorded actions and corresponding physical state changes as targets, episode-held-out splits, one predictor type for all comparisons, and positive and negative controls. Experiments on recorded robot manipulation data and a matched multi-task panel demonstrate that LASR detects known action information, rejects representative non-action shortcuts, and enables graded comparisons among imperfect latent representations. We further use LASR to train and audit a lightweight residual adapter, obtaining descriptively better point estimates whose uncertainty is assessed with paired comparisons. LASR thus provides a principled offline protocol for measuring action-related predictive support in world-model latents.
Source:https://arxiv.org/abs/2611.17775PreviousWhen Does Cross-View Consistency Matter in Multi-View Appearance Augmentation? A Controlled Study for Visuomotor PoliciesNextMultimodal Person Concept Model for Person Identification under Varying Observation Conditions
Related posts
They Found You Through a Photograph…
0dev.to
The Joy of Boring Code: Why Writing "Clever" Code is Overrated
0dev.to
From fMRI to images: what Brain-IT reconstructs—and what it does not
0dev.to
Vad ska vi automatisera först? En poängmodell för småföretag
0dev.to

Octopus Energy is bringing its sold-out plug-in solar panels to UK small businesses, says move was a 'no-brainer' following huge demand
0TechRadar
LVMH’s Media Division announces strategic partnership with OpenAI
0OpenAI Blog