SHIELD: LLM-Driven Schema Induction for Predictive Analytics in EV Battery Supply Chain Disruptions

TL;DR

SHIELD (Schema-based Hierarchical Induction for EV supply chain Disruption), a system integrating Large Language Models (LLMs) with domain expertise for EV battery supply chain risk assessment, outperforms baseline GCNs and LLM+prompt methods in disruption prediction.

Abstract

The electric vehicle (EV) battery supply chain's vulnerability to disruptions necessitates advanced predictive analytics. We present SHIELD (Schema-based Hierarchical Induction for EV supply chain Disruption), a system integrating Large Language Models (LLMs) with domain expertise for EV battery supply chain risk assessment. SHIELD combines: (1) LLM-driven schema learning to construct a comprehensive knowledge library, (2) a disruption analysis system utilizing fine-tuned language models for event extraction, multi-dimensional similarity matching for schema matching, and Graph Convolutional Networks (GCNs) with logical constraints for prediction, and (3) an interactive interface for visualizing results and incorporating expert feedback to enhance decision-making. Evaluated on 12,070 paragraphs from 365 sources (2022-2023), SHIELD outperforms baseline GCNs and LLM+prompt methods (e.g., GPT-4o) in disruption prediction. These results demonstrate SHIELD's effectiveness in combining LLM capabilities with domain expertise for enhanced supply chain risk assessment.

Figures (12)

• Figure 1: SHIELD's process for EV battery supply chain disruption prediction. The framework integrates LLM-driven schema learning with expert curation, enabling robust event extraction and prediction from diverse news sources. This approach uniquely combines LLM capabilities with domain expertise, enhancing both predictive accuracy and interpretability for proactive supply chain risk management.
• Figure 2: Overview of the supply chain schema construction process, illustrating the collection of diverse sources, schema extraction using large language models, and the integration into a unified schema library.
• Figure 3: Overview of the supply chain disruption prediction pipeline, illustrating the integration of GCN-based predictions, constrained prediction refinement, and argument coreference resolution.
• Figure 4: User interface for online disruption analysis in stage 2, showing the process from news report input to the visualization and editing of prediction results. More examples are in Appx. \ref{['app: Disruption Prediction User Interface Details']}.
• Figure 5: Sources of academic papers.