arXiv AI

GAVEL: Graph World Models for Verified and Efficient Long-Horizon LLM Task Planning

GAVEL is a framework that uses an explicit graph world model to verify and repair long‑horizon plans generated by large language models (LLMs). The graph encodes object relations, action pre‑conditions and effects, and probabilistic beliefs about unobserved object locations, allowing the system to predict action outcomes, detect violations, and repair them before execution. In experiments on BEHAVIOR‑1K, GAVEL boosts single‑task success from 41.2 % to 91.8 % and multi‑task success from 19.9 % to 92.6 %, while also reducing travel distance by about 5.4 % compared with a static variant.

arXiv Computation and Language
Oct 23

MemPlan: Memory-Conditioned PDDL Planning for Partially Observable Text Environments

MemPlan is a memory‑conditioned PDDL planning framework designed for partially observable interactive text environments. It keeps the PDDL domain fixed and reconstructs the problem at each step using parsed observations, candidate hypotheses, and interaction memory, employing positive graph memory to assign relevance‑based action costs and negative memory to record refuted assumptions as exclusion constraints. Experiments on TextWorld, ALFWorld, and Robotouille demonstrate that MemPlan improves task completion and step efficiency over language‑only and PDDL‑aware baselines while significantly reducing online token cost compared with LLM‑planning baselines such as ReAct, and its fine‑tuned interface generalizes across task types without task‑specific fine‑tuning.

By Tengwei Song, Robert Hoehndorf
arXiv AI
Sep 7

RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks?

RoboSPA is a large-scale robotic manipulation dataset and benchmark designed to evaluate Vision‑Language‑Action models on fine‑grained spatial reasoning and long‑horizon procedural planning. It contains 10 task categories, 56 base tasks, and 280 variants across five difficulty levels, with 527K trajectories collected from multiple embodiments and scenes. The benchmark introduces diagnostic metrics beyond binary success, revealing that current VLA models struggle with complex spatial relations, precise execution, and memory‑intensive planning.

By Zhenxuan Fan, Bo Zhang, Yutong Lin, Yuqian Yuan, Juekai Lin, Liang Liang, Zhuoyi Huang, Wenqiao Zhang, Juncheng Li, Siliang Tang, Jun Xiao, Yueting Zhuang
Hugging Face Trending Papers
Jul 9

GATS: Graph-Augmented Tree Search with Layered World Models for Efficient Agent Planning

Large Language Model (LLM) agents have shown promise in multi-step planning tasks, but existing approaches like LATS (Language Agent Tree Search) and ReAct rely heavily on LLM inference during planning, leading to high computational costs and stochastic behavior. We present \textbf{GATS} (Graph-Augmented Tree Search), a planning framework that combines systematic UCB1-based tree search with a layered world model to eliminate LLM calls during inference while achieving superior planning performance.

arXiv AI
Oct 9

SafeInferCom: Safe Inference-Time Compute via Verifier-Guided Mid-Generation Intervention for Robotic Task Planning

SafeInferCom introduces an inference-time monitor that verifies intermediate plans generated by large reasoning language models (LRLMs) for robotic task planning, without disrupting the decoding trajectory. The framework uses a formal verifier to preserve valid intermediate plans and guide error correction during generation, improving planning success and speeding up error correction compared to one-shot inference. Experiments across multiple LRLMs and planning domains, including VirtualHome and a real-world robotic-arm demonstration, show that SafeInferCom enhances success rates while reducing token usage when combined with iterative refinement.

By Weizhe Xu, Jialiang Fan, Mengyu Liu, Fanxin Kong