Object-Centric World Models Meet Monte Carlo Tree Search

arXiv 26.1 2026 52.1 method

TLDR

ObjectZero combines object-centric world models with GNNs and Monte Carlo Tree Search for model-based RL in dynamic environments.

Reasoning

The paper introduces a novel integration of object-level representations and GNNs with MCTS, which is a strength. However, the abstract lacks details on real-world validation and does not specify the complexity of the environment, limiting assessment of generalizability.

Read-first score

Read-first score 52.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 40.

Recency 8%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Topical relevance 42%

57.1

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=result

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 513.

Keyword Scores

world model

model-based reinforcement learning world model

world dynamics prediction

interactive world model

world simulator

generative world model

video world model

Deep Analysis

Innovations

Object-level representations for world models in RL
Use of Graph Neural Networks to model object interactions
Integration of object-centric world model with Monte Carlo Tree Search planning

Methodology

ObjectZero uses Graph Neural Networks to model interactions among multiple objects in the environment, forming an object-centric world model. This model is integrated into a model-based reinforcement learning algorithm that employs Monte Carlo Tree Search as a planning module. The algorithm is trained in a complex environment with diverse interactive objects.

Key Results

The algorithm effectively learns and predicts object dynamics in a complex setting. The results demonstrate that a structured world model with object-centric representations can be successfully integrated into model-based RL with Monte Carlo Tree Search.

Resources

Paper