arXiv:2601.21841v3 Announce Type: replace 
Abstract: While Large Language Models (LLMs) have demonstrated strong zero-shot reasoning capabilities, their deployment as embodied agents still faces fundamental challenges in long-horizon planning. Unlike open-ended text generation, embodied agents must decompose high-level intents into actionable sub-goals while adhering to the constraints of a dynamic environment. Standard LLM planners frequently fail to maintain strategy coherence over extended horizons due to context window limitations or hallucinate state transitions that violate environment constraints. We propose GiG, a planning framework that structures embodied agents' memory using a Graph-in-Graph architecture. Our approach employs a Graph Neural Network (GNN) to encode environmental states into embeddings, organizing these embeddings into action-connected execution trace graphs within an experience memory bank. GiG enables retrieval of structurally-similar priors, allowing agents to ground current decisions in relevant past structural patterns. Furthermore, we introduce a bounded lookahead module that leverages symbolic transition logic to enhance the agent's planning capabilities through grounded action projections. We evaluate our framework on three embodied planning benchmarks-Robotouille Synchronous, Robotouille Asynchronous, and ALFWorld. Our method outperforms state-of-the-art baselines, achieving Pass@1 performance gains of up to 22% on Robotouille Synchronous, 37% on Asynchronous, and 15% on ALFWorld while maintaining comparable or lower computational cost.

تم اقتراح إطار جديد يسمى GiG لتحسين قدرات التخطيط لنماذج اللغة الكبيرة (LLMs) في المهام المجسدة. يستخدم هذا الإطار بنية Graph-in-Graph لتنظيم الذاكرة وتحسين تماسك التخطيط على المدى الطويل، مما يعالج القيود التي تواجهها مخططات LLM القياسية في البيئات الديناميكية.

Se ha propuesto un nuevo marco llamado GiG para mejorar las capacidades de planificación de los Modelos de Lenguaje Grande (LLMs) en tareas encarnadas. Este marco utiliza una arquitectura Graph-in-Graph para estructurar la memoria y mejorar la coherencia de la planificación a largo plazo, abordando las limitaciones que enfrentan los planificadores LLM estándar en entornos dinámicos.

Un nouveau cadre nommé GiG a été proposé pour améliorer les capacités de planification des grands modèles de langage (LLMs) dans les tâches incarnées. Ce cadre utilise une architecture Graph-in-Graph pour structurer la mémoire et améliorer la cohérence de la planification à long terme, abordant les limitations rencontrées par les planificateurs LLM standard dans des environnements dynamiques.

A new framework named GiG has been proposed to enhance the planning capabilities of Large Language Models (LLMs) in embodied tasks. This framework utilizes a Graph-in-Graph architecture to structure memory and improve the coherence of long-horizon planning, addressing limitations faced by standard LLM planners in dynamic environments.