SSTP: Efficient Sample Selection for Trajectory Prediction

2024-09-25

AI Summary

To address the high computational cost of large-scale training and performance degradation in safety-critical scenarios (e.g., high-density traffic) caused by imbalanced scene distributions in autonomous driving trajectory prediction, this paper proposes a model-agnostic, scene-comprehensive sub-dataset construction method. It innovatively integrates gradient influence modeling with submodular function-based greedy optimization: sample gradient vectors are extracted from a pre-trained model to enable efficient and balanced sample selection. Evaluated on Argoverse 1 and 2, the method achieves full-dataset training accuracy using only 50% of the data. In high-density traffic scenarios, it reduces average displacement error (ADE) and final displacement error (FDE) by 8.2% on average, while cutting training time by 47%. The approach significantly enhances both generalization capability and operational safety without compromising prediction accuracy.

Technology Category

• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Trajectory prediction is a core task in autonomous driving. However, training advanced trajectory prediction models on large-scale datasets is both time-consuming and computationally expensive. In addition, the imbalanced distribution of driving scenarios often biases models toward data-rich cases, limiting performance in safety-critical, data-scarce conditions. To address these challenges, we propose the Sample Selection for Trajectory Prediction (SSTP) framework, which constructs a compact yet balanced dataset for trajectory prediction. SSTP consists of two main stages (1) Extraction, in which a pretrained trajectory prediction model computes gradient vectors for each sample to capture their influence on parameter updates; and (2) Selection, where a submodular function is applied to greedily choose a representative subset that covers diverse driving scenarios. This approach significantly reduces the dataset size and mitigates scenario imbalance, without sacrificing prediction accuracy and even improving in high-density cases. We evaluate our proposed SSTP on the Argoverse 1 and Argoverse 2 benchmarks using a wide range of recent state-of-the-art models. Our experiments demonstrate that SSTP achieves comparable performance to full-dataset training using only half the data while delivering substantial improvements in high-density traffic scenes and significantly reducing training time. Importantly, SSTP exhibits strong generalization and robustness, and the selected subset is model-agnostic, offering a broadly applicable solution.

Problem

• Reduces dataset size and training time for trajectory prediction
• Mitigates scenario imbalance in autonomous driving datasets
• Improves prediction accuracy in high-density traffic conditions

Innovation

• SSTP framework reduces dataset size efficiently
• Uses gradient vectors for sample influence analysis
• Applies submodular function for diverse scenario coverage