C-MAG: Cascade Multimodal Attributed Graphs for Supply Chain Link Prediction

2025-08-11

AI Summary

Addressing the challenge of precise manufacturer–product matching in global supply chains, this paper proposes C-MAG, a two-stage cascaded multimodal attribute graph model that jointly encodes textual, visual, and structural graph information to represent heterogeneous manufacturer attributes—including capabilities, certifications, and geographic constraints. Methodologically, C-MAG introduces a multi-scale message-passing mechanism and modality-aware fusion to enhance robustness under noisy conditions; it integrates heterogeneous graph neural networks, cross-modal embedding alignment, and vision-based feature extraction for end-to-end link prediction. Evaluated on the PMGraph benchmark—comprising 8,888 manufacturers and over 70,000 products—C-MAG achieves significantly higher accuracy than state-of-the-art baselines. The results demonstrate its effectiveness and practicality in improving supply chain resilience and matching efficiency.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Computer Vision: Multi-modal Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Connecting an ever-expanding catalogue of products with suitable manufacturers and suppliers is critical for resilient, efficient global supply chains, yet traditional methods struggle to capture complex capabilities, certifications, geographic constraints, and rich multimodal data of real-world manufacturer profiles. To address these gaps, we introduce PMGraph, a public benchmark of bipartite and heterogeneous multimodal supply-chain graphs linking 8,888 manufacturers, over 70k products, more than 110k manufacturer-product edges, and over 29k product images. Building on this benchmark, we propose the Cascade Multimodal Attributed Graph C-MAG, a two-stage architecture that first aligns and aggregates textual and visual attributes into intermediate group embeddings, then propagates them through a manufacturer-product hetero-graph via multiscale message passing to enhance link prediction accuracy. C-MAG also provides practical guidelines for modality-aware fusion, preserving predictive performance in noisy, real-world settings.

Problem

• Predicting manufacturer-product links in supply chains
• Handling complex multimodal manufacturer and product data
• Improving accuracy in noisy real-world supply chain scenarios

Innovation

• Aligns textual and visual attributes into embeddings
• Uses multiscale message passing in hetero-graph
• Provides modality-aware fusion for noisy data