BehaviorGPT: Smart Agent Simulation for Autonomous Driving with Next-Patch Prediction

arXiv 2024 50.2 method, application

TLDR

BehaviorGPT is a homogeneous autoregressive Transformer for traffic agent simulation using next-patch prediction, winning the Waymo Open Sim Agents Challenge.

Reasoning

The paper presents a novel approach that simplifies agent simulation by discarding the encoder-decoder architecture and history-future separation, achieving state-of-the-art results with minimal parameters. Its strength lies in the innovative NP3 paradigm and strong empirical validation on a real-world benchmark, though it is limited to autonomous driving scenarios and lacks broader world model claims.

Read-first score

Read-first score 50.2, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 41.

Recency 8%

75.1

Uses a gentle age decay so recent papers surface without erasing older foundations. 2024

Topical relevance 42%

58.6

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=none

Reproducibility 25%

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

Field roles

Candidate

Rank sensitivity

Stability: volatile; rank range: 370.

Keyword Scores

world simulator

world dynamics prediction

world model

generative world model

interactive world model

video world model

model-based reinforcement learning world model

Deep Analysis

Innovations

Homogeneous fully autoregressive Transformer that discards the history-future separation, treating each time step as the current one for motion generation, leading to a simpler and more data-efficient simulator.
Next-Patch Prediction Paradigm (NP3) that trains the model to reason at the patch level of trajectories, capturing long-range spatial-temporal interactions and mitigating the negative effects of autoregressive modeling.

Methodology

BehaviorGPT is a homogeneous autoregressive Transformer that models all time steps uniformly without distinguishing history and future. It uses the Next-Patch Prediction Paradigm (NP3) to process trajectory patches, capturing long-range spatial-temporal interactions. The model is evaluated on the Waymo Open Sim Agents Challenge, measuring realism and minADE scores.

Key Results

BehaviorGPT won first place in the 2024 Waymo Open Sim Agents Challenge with a realism score of 0.7473 and minADE of 1.4147, using only 3M parameters.

Resources

Paper