Metraipolis › archive › dc_de8dd5c8c5ea
Towards Generalizable Visually Grounded Exploration of Household Devices — arXiv:2609.00845
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.00845
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T06:32:10.138Z
- SHA-256 of the archived text
045585a53a11943400f326611642d1b966fc6876f7cc2d75314ba94a4cd7ddf4- Stable URL of this page
https://metraipolis.com/documento/dc_de8dd5c8c5ea
How to cite
Towards Generalizable Visually Grounded Exploration of Household Devices — arXiv:2609.00845 — arXiv (en), https://arxiv.org/abs/2609.00845 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_de8dd5c8c5ea (Metraipolis archive dc_de8dd5c8c5ea, sha256 045585a53a11943400f326611642d1b966fc6876f7cc2d75314ba94a4cd7ddf4)
BibTeX
@misc{dcde8dd5c8c5ea,
title = {Towards Generalizable Visually Grounded Exploration of Household Devices — arXiv:2609.00845},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.00845},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_de8dd5c8c5ea; sha256 045585a53a11943400f326611642d1b966fc6876f7cc2d75314ba94a4cd7ddf4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.00845. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recent advancements in Vision-Language Models (VLMs) have demonstrated impressive capabilities in static visual recognition and high-level semantic reasoning. However, current embodied exploration paradigms still heavily rely on imitation learning from human-annotated trajectories, which severely limits agents' generalization ability. The key bottleneck of realizing general autonomous embodied agents lies in Generalizable Visually Grounded Exploration: the ability to operate novel devices without manuals or specific training by actively grounding abstract world knowledge into fine-grained visual affordances. Yet, existing benchmarks fail to evaluate this capability: they generally rely on explicit documents and annotated trajectories, neglecting the dynamic Hypothesis-Interaction-Refinement process essential for functional device operation. To bridge this gap, we introduce VGEBench, a comprehensive benchmark designed to evaluate the generalizable visually grounded exploration capabilities of VLMs. Unlike static datasets, we construct a Logic-Driven State Machine framework. This framework simulates multi-turn interaction loops, compelling agents to achieve goals by active visual perception and feedback-driven correction. Experimental results demonstrate that existing VLMs face significant challenges in translating semantic knowledge into physical execution and maintaining long-horizon state tracking. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.00845 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T06:15:12.650Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).