Generative Active Learning for Long-tail Trajectory Prediction via Controllable Diffusion Model

2025-07-30

AI Summary

Modeling rare, long-tailed scenarios remains challenging in autonomous driving trajectory prediction due to severe data scarcity and distribution imbalance. Method: This paper proposes GALTraj, a generative active learning framework that pioneers the integration of controllable diffusion models with active learning. It introduces a tail-aware diffusion guidance mechanism that dynamically identifies error-prone tail samples during training and synthesizes traffic-plausible, diverse, and realistic rare-scenario trajectories—enabling simulation-driven long-tail learning without modifying the backbone architecture. Contribution/Results: GALTraj is plug-and-play compatible with state-of-the-art models (e.g., QCNet, MTR). Extensive evaluation on WOMD and Argoverse2 benchmarks demonstrates significant improvements in tail-scenario prediction accuracy, while simultaneously boosting head-scenario performance—achieving holistic trajectory prediction gains across the entire distribution.

Technology Category

• Machine Learning: Active Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

While data-driven trajectory prediction has enhanced the reliability of autonomous driving systems, it still struggles with rarely observed long-tail scenarios. Prior works addressed this by modifying model architectures, such as using hypernetworks. In contrast, we propose refining the training process to unlock each model's potential without altering its structure. We introduce Generative Active Learning for Trajectory prediction (GALTraj), the first method to successfully deploy generative active learning into trajectory prediction. It actively identifies rare tail samples where the model fails and augments these samples with a controllable diffusion model during training. In our framework, generating scenarios that are diverse, realistic, and preserve tail-case characteristics is paramount. Accordingly, we design a tail-aware generation method that applies tailored diffusion guidance to generate trajectories that both capture rare behaviors and respect traffic rules. Unlike prior simulation methods focused solely on scenario diversity, GALTraj is the first to show how simulator-driven augmentation benefits long-tail learning in trajectory prediction. Experiments on multiple trajectory datasets (WOMD, Argoverse2) with popular backbones (QCNet, MTR) confirm that our method significantly boosts performance on tail samples and also enhances accuracy on head samples.

Problem

• Addresses long-tail scenario challenges in trajectory prediction
• Proposes generative active learning without model architecture changes
• Enhances rare sample generation with controllable diffusion models

Innovation

• Generative active learning for trajectory prediction
• Controllable diffusion model for tail samples
• Tail-aware generation method preserves rare behaviors