How Trustworthy is Your Knowledge Graph?
summary
Shows a concrete pipeline: build a Neo4j-backed Graph RAG using AuraDB/KGBuilder and citation analysis, extract statements, add quads for provenance, then run six LLM models and apply automation plus human-in-the-loop verification. Finds six verified facts, resolves a disputed vitamin C dosage (noting a 4–10 grams range and a 10 g per-infusion citation) and catches a drug–vitamin interaction, so expect an implementation-focused demo with code/architecture and measurable results.
AI-written summary · about
chapters (7)
- 0:00Introduction to trust in AI
- 1:25Defining AI trust challenges
- 2:45Identifying verification methods
- 5:30Testing verified knowledge graphs
- 11:55Evaluating experiment results
- 20:11Architecture for prime-time use
- 24:56Conclusion and contact info
related talks
15mLightning Talk: Future-Proofing Compliance: Leveraging Knowledge Graphs and AI in Cybe...
35mGraphRag in Action: Revolutionizing Fraud Detection Techniques
34mIt's Dangerous To SLSA Alone Out There! Take This Artifact Knowledge Graph!
35mComplex Querying of Cyber Knowledge Graph for Robust Cyber Defense
24mYour company, as a Knowledge Graph - the foundation of cybersecurity’s future
28mBuilding Smarter Knowledge Graph Pipeline for Secure Semiconductor Chips
28mLLMs Change — Where Should Knowledge Live? (Lessons From SBOM)
27mChagu: Graph-Powered LLM Shielding and Anomaly Detection for Secure AI Pipelines