The current AI debate is dominated by generative models based on the statistical recognition of patterns and correlations in huge amounts of data. But while these systems excel linguistically, they lack a deeper understanding of logical relationships and causality. This lecture introduces deepassist from deepsearch – a technology that takes a radically different approach.
Instead of making decisions based on probabilities, deepassist uses an explicit world model based on a highly interconnected knowledge graph. This approach allows AI to not only recognize that something is related, but also why. By mapping conceptual causalities, AI decisions become transparent, traceable, and resilient to the typical errors of purely pattern-based systems (such as hallucinations).
The lecture highlights the architecture of this world model and discusses how the intertwining of semantic technologies and causal logic sets new standards for AI reliability in business contexts.
Literature:
Please RSVP for the onsite-lecture.
Roland Fleischhacker is the co-founder and CEO of Deepsearch GmbH, a Vienna-based software company specializing in neurosymbolic AI. He has been working in the IT industry for over 35 years. His work focuses on areas such as natural language understanding (NLU), business processes, process automation, and artificial neural networks. Recently, Roland was appointed to the University for Continuing Education Krems (UWK) to teach in the Generative AI Certificate Program.
More about Roland Fleischhacker:
Podcast Digitaler Humanismus: Korrelation vs. Kausalität - Roland Fleischhacker über die Grenzen und Potenziale der KI (2024)
17 March 2026
16:00-18:00
Österreichische Akademie der Wissenschaften, Bäckerstraße 13, 1010 Wien, Seminarraum 1, Erdgeschoss / Innenhof
English
