⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

Temporal Triplane Transformers as Occupancy World Models

arXiv 25.3 2025 52.4 method

TLDR

A novel 4D occupancy world model for autonomous driving using triplane representations and incremental delta prediction, achieving faster and more accurate planning.

Reasoning

The paper introduces a compact triplane representation and incremental prediction strategy, which are novel and effective, as shown by improved speed and accuracy. However, the abstract lacks details on real-world dataset specifics and comparisons to other world model types, limiting generalizability assessment.

Read-first score

Read-first score 52.4, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 39.

Recency 8%
86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Topical relevance 42%
55.7

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%
50

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

Reproducibility 25%
38

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=code

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 417.

Keyword Scores

world model
10
world dynamics prediction
9
generative world model
6
world simulator
4
interactive world model
4
video world model
3
model-based reinforcement learning world model
3

Deep Analysis

Innovations

  • Triplane-based representation for compact 3D occupancy encoding
  • Incremental prediction strategy that models changes in occupancy (delta prediction) rather than full states

Methodology

Delta-Triplane Transformers (DTT) encode 3D occupancy into a compact triplane representation, extract multi-scale motion features from historical data, and iteratively predict future triplane deltas. These deltas are combined with past states to decode future occupancy and ego-motion trajectories.

Key Results

DTT achieves a 1.44× speedup (26 FPS) over the state of the art, improves mean IoU to 30.85, and reduces mean absolute planning error to 1.0 meters.

Resources

Paper

Tags