MetaDAT: Generalizable Trajectory Prediction via Meta Pre-training and Data-Adaptive Test-Time Updating

2026-03-10

AI Summary

This work addresses the performance degradation of trajectory prediction models during testing due to distribution shifts by proposing a meta-learning pretraining framework tailored for trajectory forecasting, which integrates test-time training (TTT) with a data-adaptive updating mechanism. The approach employs bi-level optimization during meta-pretraining to enhance the model’s rapid adaptation capability. At test time, it dynamically adjusts the learning rate and update frequency, while leveraging hard example mining and online gradient analysis to focus on critical samples, enabling efficient and accurate online adaptation. Evaluated across diverse datasets—including nuScenes, Lyft, and Waymo—the method significantly outperforms existing approaches and demonstrates robustness and high performance even under suboptimal learning rates or high frame rates.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Online Learning & Bandits
• 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
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Existing trajectory prediction methods exhibit significant performance degradation under distribution shifts during test time. Although test-time training techniques have been explored to enable adaptation, current approaches rely on an offline pre-trained predictor that lacks online learning flexibility. Moreover, they depend on fixed online model updating rules that do not accommodate the specific characteristics of test data. To address these limitations, we first propose a meta-learning framework to directly optimize the predictor for fast and accurate online adaptation, which performs bi-level optimization on the performance of simulated test-time adaptation tasks during pre-training. Furthermore, at test time, we introduce a data-adaptive model updating mechanism that dynamically adjusts the predefined learning rates and updating frequencies based on online partial derivatives and hard sample selection. This mechanism enables the online learning rate to suit the test data, and focuses on informative hard samples to enhance efficiency. Experiments are conducted on various challenging cross-dataset distribution shift scenarios, including nuScenes, Lyft, and Waymo. Results demonstrate that our method achieves superior adaptation accuracy, surpassing state-of-the-art test-time training methods for trajectory prediction. Additionally, our method excels under suboptimal learning rates and high FPS demands, showcasing its robustness and practicality.

Problem

• trajectory prediction
• distribution shift
• test-time adaptation
• online learning
• model updating

Innovation

• meta-learning
• test-time adaptation
• trajectory prediction
• data-adaptive updating
• distribution shift