HGOE: Hybrid External and Internal Graph Outlier Exposure for Graph Out-of-Distribution Detection
Abstract: With the progressive advancements in deep graph learning, out-of-distribution (OOD) detection for graph data has emerged as a critical challenge. While the efficacy of auxiliary datasets in enhancing OOD detection has been extensively studied for image and text data, such approaches have not yet been explored for graph data. Unlike Euclidean data, graph data exhibits greater diversity but lower robustness to perturbations, complicating the integration of outliers. To tackle these challenges, we propose the introduction of \textbf{H}ybrid External and Internal \textbf{G}raph \textbf{O}utlier \textbf{E}xposure (HGOE) to improve graph OOD detection performance. Our framework involves using realistic external graph data from various domains and synthesizing internal outliers within ID subgroups to address the poor robustness and presence of OOD samples within the ID class. Furthermore, we develop a boundary-aware OE loss that adaptively assigns weights to outliers, maximizing the use of high-quality OOD samples while minimizing the impact of low-quality ones. Our proposed HGOE framework is model-agnostic and designed to enhance the effectiveness of existing graph OOD detection models. Experimental results demonstrate that our HGOE framework can significantly improve the performance of existing OOD detection models across all 8 real datasets.
Primary Subject Area: [Content] Multimodal Fusion
Secondary Subject Area: [Experience] Multimedia Applications
Relevance To Conference: The proposed Hybrid External and Internal Graph Outlier Exposure (HGOE) framework is highly relevant to the conference's focus on Multimodal Fusion and Multimedia Applications. HGOE addresses the critical challenge of out-of-distribution detection in graph data, which is essential for representing complex relationships within multimedia data, such as images, text, and videos. By enhancing the performance of existing graph OOD detection models, HGOE plays a role in facilitating better multimodal fusion and understanding of multimedia data. This improved performance indirectly supports multimedia applications, such as social networks and recommendation systems, which rely on graph representations for accurate and context-aware interactions. As a model-agnostic approach, HGOE seamlessly integrates with various multimedia processing tasks, promoting advancements in multimodal fusion techniques and the reliability of multimedia applications. The experimental results on real datasets further demonstrate the relevance of this work to the conference's themes and its potential impact on the field.
Submission Number: 2652
Loading