About

CAOS: Cognition And OntologieS, celebrates its 10th edition!
CAOS investigates fundamental cognitive phenomena and concepts across language, psychology, and reasoning, examining how these can be formally and ontologically analyzed with the purpose to model, simulate and represent cognitive phenomena for artificial intelligence in the context of both traditional symbolic AI and contemporary neural approaches.

In 2026, CAOS 10 is organised at special second edition of TriCoLore held in Bressanone/Brixen, Italy (1-6 December 2026).

Claims regarding Artificial General Intelligence (AGI) and its purported present-day manifestation are becoming increasingly prevalent. In contrast, skeptics of the Large Language Models (LLM) approach characterize current models as examples of Broad, Shallow Intelligence (BSI). The artificial intelligence community’s discourse appears to be bifurcated: one aspect focuses on the reconceptualization of “intelligence,” while another emphasizes a functional approach wherein embodied cognition, cognitive computing, and the intersection of symbolic and neuro-symbolic methodologies with cognitive sciences have become crucial considerations.

Simultaneously, conceptual modeling is gaining prominence in corporate environments too, where “the O-word”, namely “Ontology”, is not a taboo anymore, and is now embraced. The significance and advantages of a hybrid methodology that integrates ontologies, knowledge graphs, and LLMs, coupled with cognitively-informed data model conceptualization, have become increasingly apparent.The capacity for abstraction and deterministic, complex aggregate cognition appears to currently remain a human advantage over generative AI. Emerging frontiers in robotics raise fundamental questions about embodied cognition and cognitive computing. The convergence of knowledge representation, symbolic approaches, and LLM capabilities presents substantial opportunities for advancing our understanding of cognitive phenomena. Furthermore, the evolution of artificial intelligence and our comprehension of cognitive phenomena can be substantially enhanced through traditional methodologies in knowledge representation and ontologies.

CAOS endeavors to bridge the divide between cognitive science and formal methods by establishing a forum for researchers from both domains to present and discuss their work.

Cover image credits to Oliver Kutz.