REVIEW 8 cited by
WALL-E: World Alignment by Rule Learning Improves World Model-based LLM Agents
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
WALL-E: World Alignment by Rule Learning Improves World Model-based LLM Agents
read the original abstract
Can large language models (LLMs) directly serve as powerful world models for model-based agents? While the gaps between the prior knowledge of LLMs and the specified environment's dynamics do exist, our study reveals that the gaps can be bridged by aligning an LLM with its deployed environment and such "world alignment" can be efficiently achieved by rule learning on LLMs. Given the rich prior knowledge of LLMs, only a few additional rules suffice to align LLM predictions with the specified environment dynamics. To this end, we propose a neurosymbolic approach to learn these rules gradient-free through LLMs, by inducing, updating, and pruning rules based on comparisons of agent-explored trajectories and world model predictions. The resulting world model is composed of the LLM and the learned rules. Our embodied LLM agent "WALL-E" is built upon model-predictive control (MPC). By optimizing look-ahead actions based on the precise world model, MPC significantly improves exploration and learning efficiency. Compared to existing LLM agents, WALL-E's reasoning only requires a few principal rules rather than verbose buffered trajectories being included in the LLM input. On open-world challenges in Minecraft and ALFWorld, WALL-E achieves higher success rates than existing methods, with lower costs on replanning time and the number of tokens used for reasoning. In Minecraft, WALL-E exceeds baselines by 15-30% in success rate while costing 8-20 fewer replanning rounds and only 60-80% of tokens. In ALFWorld, its success rate surges to a new record high of 95% only after 6 iterations.
Forward citations
Cited by 8 Pith papers
-
Agent vs. Parametric World Models: Hybrid Planning for Reliable Language Agents
GILP trains a parameterized backbone for valid actions and state predictions, then uses a consistency gate with LLM drafts to reduce hallucinated-state rate from 0.176 to 0.035 on GPT-4o-mini while raising success fro...
-
Beyond Next-Observation Prediction: Agent-Authored World Modeling for Sequential Decision Making
AAWM builds training targets for world models by retrieving and synthesizing transition evidence based on the policy's self-identified decision needs at each state.
-
Object-Centric Environment Modeling for Agentic Tasks
Object-Centric Environment Modeling (OCM) builds an online executable object-and-procedure code model that improves average rank and cuts invalid actions on ScienceWorld, ALFWorld, and PlanCraft.
-
Agent vs. Parametric World Models: Hybrid Planning for Reliable Language Agents
GILP combines a small parameterized world model with LLM agent reasoning via a consistency gate, reducing hallucinated-state rate from 0.176 to 0.035 and raising success from 0.668 to 0.838 on graph planning benchmarks.
-
Agent vs. Parametric World Models: Hybrid Planning for Reliable Language Agents
A small parametric transition model plus a Jaccard consistency gate grounds LLM agent state deltas, cutting hallucinated-state rate ~80% and raising success from 0.668 to 0.838 on graph planning benchmarks.
-
Mini Amusement Parks (MAPs): A Testbed for Modelling Business Decisions
MAPs is a new amusement-park simulator benchmark on which frontier LLM agents score 7–15% of human performance, exposing persistent gaps in long-horizon planning, active learning, spatial reasoning, and handling stoch...
-
ReCAPA: Hierarchical Predictive Correction to Mitigate Cascading Failures
ReCAPA adds predictive correction and multi-level semantic alignment to VLA models, plus two new metrics for tracking error spread and recovery, yielding competitive benchmark results over LLM baselines.
-
ReCAPA: Hierarchical Predictive Correction to Mitigate Cascading Failures
ReCAPA uses multi-level predictive correction and semantic alignment modules to reduce cascading failures in VLA systems, with new metrics for tracking error propagation and recovery on embodied benchmarks.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.