Скачать или смотреть Powering Legal Research with Knowledge Graphs and Large Language Models

The EU competition case database serves as a key reference for competition law experts at the Central Bank of Egypt. However, in its current linear format, it offers limited metadata-based search, making it challenging to uncover complex patterns such as recurring legal grounds, overlapping corporate entities, or coordinated criminal networks. Basant will open the session by briefly explaining the workflow of competition law research and setting the context for this problem.

To overcome these challenges, the data science team ingested open-source EU Commission cases spanning five domains: mergers, state aid, antitrust and cartels, the Digital Markets Act, and foreign subsidies regulation. Ahmad will present the technical implementation, highlighting how the team: (1) modeled an ontology capturing both intracase and intercase links from the case texts, (2) constructed a legal knowledge graph using Neo4j with the aid of an LLM, and (3) used Neo4j Bloom as an interface to interact with the knowledge graph.

The resulting tool empowers competition law experts to search using near natural language to visualize subgraphs representing specific legal perspectives. Such insight informs their research, ultimately leading to proactive financial market regulation. Dr Asmaa will conclude the session with a live demonstration of the product.

Speakers: Basant Mounir & Ahmad Abdelaal
View Presentation: https://docs.google.com/presentation/...

Resources:
Get Started with Aura - https://bit.ly/3LOLrjh
Deployment Center - https://bit.ly/4jOelM3
Ground AI Systems and Agents with Neo4j - https://bit.ly/4oVsnyb


#nodes2025 #neo4j #graphdatabase #graphrag #knowledgegraph