Talk to your data: leverage open schema and knowledge for efficient self-service insights
A Talk by Atanas Kiryakov (CEO & Founder, Ontotext)
About this Talk
Discover how to democratize data access and eliminate the traditional bottleneck where business users must wait weeks for technical experts to build queries. This technical presentation explores cutting-edge approaches to self-service analytics using knowledge graphs, semantic technologies, and generative AI.
Key Topics Covered
GraphRAG Architecture Patterns
Learn about three distinct GraphRAG implementations:
- Type 1: Content metadata store with enhanced filtering capabilities
- Type 2: Domain knowledge enrichment for contextual understanding
- Type 3: Direct graph querying with natural language to SPARQL translation
Overcoming Traditional Knowledge Graph Limitations
Address the common challenge where sophisticated knowledge graphs become inaccessible to non-technical users, requiring specialized SPARQL expertise and creating organizational data silos.
Leveraging Open Standards for AI Integration
Explore how using established ontologies like Schema.org and existing knowledge bases such as Wikidata dramatically reduces implementation complexity. Large language models are already trained on these standards, eliminating the need for custom fine-tuning.
Real-World Implementation: The NASA Case Study
Follow a practical demonstration using 1,000 marketing documents to solve entity disambiguation challenges. See how traditional chunking-based RAG fails to connect implicit knowledge (NASA as a government organization) and how graph-enhanced approaches succeed.
Technical Architecture Deep Dive
- Multi-modal data fabric combining structured data, documents, and metadata
- 12+ billion triple knowledge graph implementation
- Entity linking services for automatic knowledge graph population
- Self-healing SPARQL query generation with automatic error correction
Development Efficiency Breakthrough
Witness how modern GraphRAG tooling reduces implementation time from days/weeks to hours, with live demonstrations of natural language querying against massive knowledge graphs.
Performance Optimization Strategies
Understand the four critical subtasks in GraphRAG implementations:
- Relevant chunk retrieval
- Entity extraction and linking
- Database querying
- Response synthesis
Learn when to use LLMs versus specialized models for optimal speed, accuracy, and cost-effectiveness.
Target Audience
- Data scientists and engineers
- Solution architects
- Business intelligence professionals
- Knowledge management specialists
- Organizations seeking to democratize data access
Key Takeaways
- Transform knowledge graphs from expert-only tools to self-service platforms
- Leverage existing ontologies and standards for faster AI integration
- Implement GraphRAG patterns appropriate for your data architecture
- Achieve enterprise-scale semantic search with minimal development overhead
- Bridge the gap between complex semantic technologies and business user needs
This presentation combines 25+ years of semantic web expertise with practical AI implementation strategies, offering attendees actionable insights for building next-generation self-service analytics platforms using knowledge graphs and generative AI.
About The Speaker
We’re a cloud tech company that provides organisations around the world with computing infrastructure and software to help them innovate, unlock efficiencies and become more effective. We also created the world’s first – and only – autonomous database to help organise and secure our customers’ data.
Since 2016 Neural Alpha have delivered cutting edge, sustainability centric Connected Data solutions for blue-chip corporates, financial institutions, Governments and NGOs. Our bespoke software & data solutions fuse AI, Knowledge Graphs, Taxonomies & other technologies for unprecedented insights.