ESCA: Contextualizing Embodied Agents via Scene-Graph Generation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Jiani, Sethi, Amish, Kuo, Matthew, Keoliya, Mayank, Velingker, Neelay, Jung, JungHo, Lim, Ser-Nam, Li, Ziyang, Naik, Mayur
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911233560543232
author Huang, Jiani
Sethi, Amish
Kuo, Matthew
Keoliya, Mayank
Velingker, Neelay
Jung, JungHo
Lim, Ser-Nam
Li, Ziyang
Naik, Mayur
author_facet Huang, Jiani
Sethi, Amish
Kuo, Matthew
Keoliya, Mayank
Velingker, Neelay
Jung, JungHo
Lim, Ser-Nam
Li, Ziyang
Naik, Mayur
contents Multi-modal large language models (MLLMs) are making rapid progress toward general-purpose embodied agents. However, existing MLLMs do not reliably capture fine-grained links between low-level visual features and high-level textual semantics, leading to weak grounding and inaccurate perception. To overcome this challenge, we propose ESCA, a framework that contextualizes embodied agents by grounding their perception in spatial-temporal scene graphs. At its core is SGCLIP, a novel, open-domain, promptable foundation model for generating scene graphs that is based on CLIP. SGCLIP is trained on 87K+ open-domain videos using a neurosymbolic pipeline that aligns automatically generated captions with scene graphs produced by the model itself, eliminating the need for human-labeled annotations. We demonstrate that SGCLIP excels in both prompt-based inference and task-specific fine-tuning, achieving state-of-the-art results on scene graph generation and action localization benchmarks. ESCA with SGCLIP improves perception for embodied agents based on both open-source and commercial MLLMs, achieving state of-the-art performance across two embodied environments. Notably, ESCA significantly reduces agent perception errors and enables open-source models to surpass proprietary baselines. We release the source code for SGCLIP model training at https://github.com/video-fm/LASER and for the embodied agent at https://github.com/video-fm/ESCA.
format Preprint
id arxiv_https___arxiv_org_abs_2510_15963
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ESCA: Contextualizing Embodied Agents via Scene-Graph Generation
Huang, Jiani
Sethi, Amish
Kuo, Matthew
Keoliya, Mayank
Velingker, Neelay
Jung, JungHo
Lim, Ser-Nam
Li, Ziyang
Naik, Mayur
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Multi-modal large language models (MLLMs) are making rapid progress toward general-purpose embodied agents. However, existing MLLMs do not reliably capture fine-grained links between low-level visual features and high-level textual semantics, leading to weak grounding and inaccurate perception. To overcome this challenge, we propose ESCA, a framework that contextualizes embodied agents by grounding their perception in spatial-temporal scene graphs. At its core is SGCLIP, a novel, open-domain, promptable foundation model for generating scene graphs that is based on CLIP. SGCLIP is trained on 87K+ open-domain videos using a neurosymbolic pipeline that aligns automatically generated captions with scene graphs produced by the model itself, eliminating the need for human-labeled annotations. We demonstrate that SGCLIP excels in both prompt-based inference and task-specific fine-tuning, achieving state-of-the-art results on scene graph generation and action localization benchmarks. ESCA with SGCLIP improves perception for embodied agents based on both open-source and commercial MLLMs, achieving state of-the-art performance across two embodied environments. Notably, ESCA significantly reduces agent perception errors and enables open-source models to surpass proprietary baselines. We release the source code for SGCLIP model training at https://github.com/video-fm/LASER and for the embodied agent at https://github.com/video-fm/ESCA.
title ESCA: Contextualizing Embodied Agents via Scene-Graph Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2510.15963