https://www.deeplearning.ai/short-courses/agentic-knowledge-graph-construction/
Course Overview
DeepLearning.AI has released a new free short course on Agentic Knowledge Graph Construction, taught by Andreas Kollegger from Neo4j.
This course teaches how to automate the construction of knowledge graphs using collaborative agents, moving beyond traditional RAG systems to better represent relationships within your data.
What You’ll Learn
- Design specialized agents that identify user goals and suggest nodes/relationships to extract from structured and unstructured data
- Implement multi-agent systems using Google’s Agent Development Kit (ADK) to orchestrate specialized agents
- Construct knowledge graphs based on proposed schemas and connect them into complete graphs
Course Details
Duration: 12 lessons with 8 code examples
Instructor: Andreas Kollegger, Developer Evangelist for Generative AI at Neo4j
Prerequisites: Python familiarity (Cypher knowledge helpful but not required)
Cost: Free
Key Topics Covered
- Understanding knowledge graphs and how they capture relationships
- Multi-agent system architecture with conversational agents and sub-agentic workflows
- Using Google’s ADK to build and coordinate agents with shared context
- Building user intent agents and file-suggestion agents
- Creating graph schemas from both structured (CSV) and unstructured (text) data
- Connecting multiple graphs into unified knowledge graphs
Practical Application
The course includes a hands-on project where you’ll build an agentic system to find root causes of product issues using:
- Structured data: Product and supplier information
- Unstructured data: Product reviews
- Graph relationships: Products, parts, suppliers, and issues
Who Should Take This Course?
- Developers working with RAG systems wanting to improve retrieval accuracy
- Those curious about agentic design patterns
- Anyone interested in knowledge graph databases
Source: DeepLearning.AI Short Courses