ARGUS: Role-Aware Event Knowledge Graphs for U.S. Employment-Discrimination Complaints
ARGUS builds event knowledge graphs from U.S. employment-discrimination complaints to improve claim classification and legal QA.
ARGUS is a source-grounded pipeline that builds document-level event knowledge graphs from CourtListener U.S. employment-discrimination complaints. It combines a 5W1H-inspired schema, legal-domain models, and LLM structured generation to capture participants, time, and causal links, then merges chunk graphs into document graphs. A graph-structured classifier outperformed raw and linearized baselines on held-out claim classification, and EKG-only retrieval improved document-scoped legal QA, while open-retrieval gains were limited by low first-stage recall.
- ARGUS builds document-level event graphs from discrimination complaints.
- It combines a 5W1H schema, legal models, and LLM generation.
- A graph classifier beat raw and linearized baselines on claim classification.
- EKG retrieval improved document-scoped QA; open retrieval stayed recall-limited.
Full article132 words · extracted from arxiv.org · click to collapse
U.S. employment-discrimination complaints describe complex event sequences that are not explicitly captured by lexical or embedding-based representations alone. We present ARGUS, a source-grounded pipeline that combines a 5W1H-inspired schema, legal-domain models, and LLM-based structured generation to construct document-level Event Knowledge Graphs (EKGs) from CourtListener complaints. ARGUS extracts fact-bearing statements, builds chunk-level event graphs with participant, temporal, and causal structure, and merges them into document-level representations. We evaluate graph quality through human and multi-model assessment and test downstream utility on claim classification and legal QA. The graph-structured classifier outperforms raw and linearized baselines on the held-out set, and EKG-only retrieval improves document-scoped QA, while open-retrieval gains remain limited by low first-stage candidate recall. These results suggest that EKGs are most useful for organizing and reasoning over evidence once relevant material has been retrieved.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.30184