ST-ProC: A Graph-Prototypical Framework for Robust Semi-Supervised Travel Mode Identification

2025-11-17

AI Summary

GPS trajectory mode identification (TMI) suffers from high annotation costs and extreme label scarcity. Existing semi-supervised learning (SSL) methods are prone to catastrophic confirmation bias and neglect the intrinsic manifold structure of trajectory data. To address these issues, we propose a robust semi-supervised framework integrating manifold-aware representation learning and prototype-based classification. Specifically, we employ a graph neural network to explicitly model the trajectory manifold; introduce graph-prototype centroids as class-specific anchors; design a margin-aware pseudo-labeling strategy to suppress erroneous labeling; and jointly optimize via graph regularization and teacher-student consistency. This approach effectively mitigates confirmation bias and enhances pseudo-label reliability. Under realistic sparse-labeling settings, our method achieves a 21.5% accuracy improvement over state-of-the-art SSL baselines such as FixMatch.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning/Scheduling and Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Travel mode identification (TMI) from GPS trajectories is critical for urban intelligence, but is hampered by the high cost of annotation, leading to severe label scarcity. Prevailing semi-supervised learning (SSL) methods are ill-suited for this task, as they suffer from catastrophic confirmation bias and ignore the intrinsic data manifold. We propose ST-ProC, a novel graph-prototypical multi-objective SSL framework to address these limitations. Our framework synergizes a graph-prototypical core with foundational SSL Support. The core exploits the data manifold via graph regularization, prototypical anchoring, and a novel, margin-aware pseudo-labeling strategy to actively reject noise. This core is supported and stabilized by foundational contrastive and teacher-student consistency losses, ensuring high-quality representations and robust optimization. ST-ProC outperforms all baselines by a significant margin, demonstrating its efficacy in real-world sparse-label settings, with a performance boost of 21.5% over state-of-the-art methods like FixMatch.

Problem

• Addresses label scarcity in GPS-based travel mode identification
• Overcomes confirmation bias and manifold neglect in semi-supervised learning
• Improves robustness through graph regularization and noise-resistant pseudo-labeling

Innovation

• Graph-prototypical SSL framework for travel mode identification
• Margin-aware pseudo-labeling strategy to actively reject noise
• Synergy of contrastive and teacher-student consistency losses