OT-AIAT 09
 Knowledge Bases for AI Agents

Zielgruppe

The webinar is aimed at everyone who wants to understand knowledge graphs and how to combine them with AI agents.

Attendance is open and free of charge for all participants from academia, industry, and public administration fromEUand/orEuroHPC JUmember countries.

Entry level & prerequisites

Basic - no prior experience required.

Inhalt und Ziele

Learn how to supercharge AI agents with structured knowledge bases and knowledge graphs. This training explores how to build, connect, and query knowledge representations that give LLM-based agents reliable, up-to-date, and domain-specific information for smarter decision-making.

LLMs are powerful, but on their own they often struggle with domain-specific, up-to-date, or structured knowledge - leading to hallucinations, unreliable and contextually irrelevant outputs. This training addresses a key challenge in building effective AI agents: equipping them with external knowledge bases that provide accurate, structured, and queryable information. Participants will explore how knowledge graphs and knowledge bases can serve as a persistent „memory“ and factual backbone for LLM-based agents, enabling them to reason over real-world entities, relationships, and domain-specific facts rather than relying solely on pre-trained general knowledge.

The session covers the spectrum from lightweight approaches - such as connecting agents to curated wikis and document stores via Retrieval-Augmented Generation (RAG) - to more advanced architectures that integrate knowledge graphs with graph-based querying (e.g., SPARQL, Cypher) to enable structured reasoning. Participants will learn how domain knowledge is modelled as a graph, populate it from various data sources, and integrate it with AI agents so that the agent can look up facts, traverse relationships, and ground its responses in verified information. Whether you are building internal company knowledge assistants, domain-specific Q&A systems, or AI agents to work with your Personal Knowledge Graph, this training provides the practical patterns to get started.

Agenda

  • Welcome & Introduction
  • Knowledge Graphs: Concepts, Entities & Relationships
  • AI Agents with Knowledge Bases: Connecting LLMs to Obsidian-style Wikis & Document Stores
  • Break
  • AI Agents with Knowledge Graphs: Traversal, Multi-hop Queries & Grounded Reasoning
  • Choosing the Right Architecture for Your Use Case
  • Questions & Discussion

Speakers

Daniel Dobriy (WU Vienna, Bilateral AI Cluster of Excellence, Dobriy AI GmbH)
Daniel Dobriy is an AI researcher and lecturer at the Institute for Data, Process and Knowledge Management at WU Vienna. His work bridges symbolic AI, graph-based machine learning, and applications in public governance, finance, education, and scientific discovery. He is a Research Fellow at the Bilateral AI Cluster of Excellence, an industry leader in the GOBLIN COST Action, and an active member of the W3C RDF and SPARQL Working Group. Daniel Dobriy is certified in Value-based Engineering ISO 24748-7000, is a member of ASAI, ÖCG, ACM and ACL, chairs RAGE-KG and serves on programme committees of leading international venues including The Web Conference, ISWC (International Semantic Web Conference), ESWC, K-CAP, HICSS, SEMANTiCS as well as IJCKG. Daniel Dobriy is also the managing director of Dobriy AI GmbH (dobriy.ai).

Trainer

Zeit und Ort

Seminarnummer:
OT-AIAT 09
Trainingsform:
Online-Tr. / Webinar
Dauer:
1 Tag(e) (2 Stunden)
Termin:
03.09.2026
Stundenplan:
16:00 - 18:00
Ort:
Online

Anmerkungen

This course will be delivered as an ONLINE Course via ZOOM.

Organisation

Seminarleitung:
Buchwinkler David BA MA
Sekretariat:
Fröhlich Elisabeth