Combining Semantic Embeddings and Knowledge Graphs for Identifying Decision Patterns in Brazilian Judicial Decisions

2026-04-12 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Legal AI Applications · Depth: Expert, short

Summary

Researchers Gustavo Soares Silva, Omar Andres Carmona Cortes, Fábio Manoel França Lobato, and Antonio Fernando Lavareda Jacob Junior are developing a dynamic clustering system for Brazilian judicial decisions. This system addresses limitations of text-only approaches by integrating hybrid representations, specifically combining semantic embeddings from legal-domain Portuguese models with knowledge graphs automatically constructed from the documents. The architecture is designed to support incremental clustering, allowing for continuous updates, and uses Large Language Models (LLMs) to generate justifications for clusters based on the relations within the knowledge graphs. Preliminary evaluations of this system utilize quantitative metrics such as the Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index to assess its performance.

Key takeaway

For legal tech developers building decision support systems, integrating hybrid representations of semantic embeddings and knowledge graphs can significantly improve the identification of complex patterns in judicial texts. You should consider this approach to overcome the limitations of purely textual methods, especially for documents with high lexical similarity, and to enable more robust, explainable clustering of legal data.

Key insights

Hybrid representations combining semantic embeddings and knowledge graphs enhance judicial decision pattern identification.

Principles

• Text-only approaches limit structural relation capture.
• Knowledge graphs ground LLM justifications.

Method

The system uses legal-domain Portuguese semantic embeddings and automatically constructed knowledge graphs for dynamic, incremental clustering of judicial decisions, with LLMs providing cluster justifications.

In practice

• Apply hybrid representations for legal text analysis.
• Use LLMs for explainable clustering justifications.

Topics

• Judicial Decision Analysis
• Semantic Embeddings
• Knowledge Graphs
• Dynamic Clustering
• Legal-domain Portuguese Models

Best for: Research Scientist, AI Scientist, NLP Engineer, Legal Professional

Related on AIssential

• A Hybrid Framework for Natural Language Querying of IFC Models with Relational and Graph Representations — Combining relational and graph representations with iterative LLM reasoning enhances IFC model accessibility.
• HKJudge: A Legal Discourse-Annotated Corpus for Interpreting What Courts Find, How They Reason, and What They Rule — Expert-annotated legal discourse corpora enable advanced AI understanding of court judgments.
• Retrieval-Augmented Generation and Knowledge Graphs in Portuguese-Language Legal Documents — GraphRAG enhances legal QA by modeling document structure as knowledge graphs for context-aware retrieval.
• Democratizing Legal Analytics: Resource-Efficient Information Extraction for Brazilian Case Law — Resource-efficient LLM fine-tuning enables scalable legal analytics in low-resource languages using legacy datasets.
• TW-LegalBench: Measuring Taiwanese Legal Understanding — LLMs demonstrate proficiency in legal qualification exams but struggle with precise statutory citation and advanced reasoning in civil-law contexts.
• Agents + Knowledge Graphs with NetDocuments’ Dan Hauck — Knowledge graphs enhance legal AI agent performance by providing crucial document context.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.