Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lee, Xian Yeow, Vidyaratne, Lasitha, Sin, Gregory, Farahat, Ahmed, Gupta, Chetan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2511.18319
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911282940084224
author Lee, Xian Yeow
Vidyaratne, Lasitha
Sin, Gregory
Farahat, Ahmed
Gupta, Chetan
author_facet Lee, Xian Yeow
Vidyaratne, Lasitha
Sin, Gregory
Farahat, Ahmed
Gupta, Chetan
contents Autonomous inspection in hazardous environments requires AI agents that can interpret high-level goals and execute precise control. A key capability for such agents is spatial grounding, for example when a drone must center a detected object in its camera view to enable reliable inspection. While large language models provide a natural interface for specifying goals, using them directly for visual control achieves only 58\% success in this task. We envision that equipping agents with a world model as a tool would allow them to roll out candidate actions and perform better in spatially grounded settings, but conventional world models are data and compute intensive. To address this, we propose a task-specific latent dynamics model that learns state-specific action-induced shifts in a shared latent space using only goal-state supervision. The model leverages global action embeddings and complementary training losses to stabilize learning. In experiments, our approach achieves 71\% success and generalizes to unseen images and instructions, highlighting the potential of compact, domain-specific latent dynamics models for spatial alignment in autonomous inspection.
format Preprint
id arxiv_https___arxiv_org_abs_2511_18319
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Weakly-supervised Latent Models for Task-specific Visual-Language Control
Lee, Xian Yeow
Vidyaratne, Lasitha
Sin, Gregory
Farahat, Ahmed
Gupta, Chetan
Artificial Intelligence
Machine Learning
Systems and Control
Autonomous inspection in hazardous environments requires AI agents that can interpret high-level goals and execute precise control. A key capability for such agents is spatial grounding, for example when a drone must center a detected object in its camera view to enable reliable inspection. While large language models provide a natural interface for specifying goals, using them directly for visual control achieves only 58\% success in this task. We envision that equipping agents with a world model as a tool would allow them to roll out candidate actions and perform better in spatially grounded settings, but conventional world models are data and compute intensive. To address this, we propose a task-specific latent dynamics model that learns state-specific action-induced shifts in a shared latent space using only goal-state supervision. The model leverages global action embeddings and complementary training losses to stabilize learning. In experiments, our approach achieves 71\% success and generalizes to unseen images and instructions, highlighting the potential of compact, domain-specific latent dynamics models for spatial alignment in autonomous inspection.
title Weakly-supervised Latent Models for Task-specific Visual-Language Control
topic Artificial Intelligence
Machine Learning
Systems and Control
url https://arxiv.org/abs/2511.18319