Streamlining Legislative Compliance with Neo4j
summary
Capgemini + Neo4j pipeline ingests the DORA PDF, cleans text with regex, chunks at ~1024 chars, runs SpaCy coreference resolution, NER and relation extraction, creates HuggingFace Llama embeddings and stores them in a Neo4j hybrid vector+graph database, then answers RAG queries via a small Flask frontend. Uses a local Llama 3.1 model for generation. Outputs: concise regulation summaries, applicability checks (jurisdiction, revenue thresholds), affected departments, and concrete DORA requirements like due diligence, conflict assessment and audit rights.
AI-written summary · about
chapters (8)
- 0:00Introduction to talk
- 1:16Project vision and goals
- 3:03DORA regulation overview
- 4:47Processing regulation documents
- 6:58Coreference resolution explained
- 8:01Named entity recognition
- 8:45Extracting entity relationships
- 9:49Demo and Q&A examples
related talks
36mGraphRAG Explained 🔥 Building Smarter LLMs
16mGraph-Powered Compliance: Building Intelligent Regulatory Mapping with AI and Neo4j
30mFrom Unstructured Docs to Graph Intelligence: A Framework for Building Graph+Embedding Pipelines
15mAgentic GraphRAG: Simplifying Retrieval Across Structured & Unstructured Data
36mGraphRAG for Law: Building Legal Reasoning Agents
15mGraphRAG-Powered AI Agents: Transforming Corporate Metadata into Intelligent Knowledge
29mPowering Legal Research with Knowledge Graphs and Large Language Models
27mGraphReader Agentic RAG: Rethinking Long-Context Retrieval Systems