Computing Meaning from Conceptual Structures in Integrated Information Theory

Explore a groundbreaking approach to computing meaning within the framework of Integrated Information Theory (IIT) in this 30-minute conference talk. Delve into the postulation that conscious experience generates not only information but also intrinsic, compositional, specific, and integrated meaning. Discover how the category theoretic formulation of Distributional Semantics can be applied to compute the meaning of a Q-shape, or maximally irreducible conceptual structure in IIT. Learn about the assignment of distributional meanings and grammatical types to concepts, the identification of core concepts, and the use of vector spaces to represent conceptual relationships. Examine the application of pre-group algebra for ordering grammatical types and the utilization of monoidal categories to compute compositional meaning. Gain insights from a simple toy model demonstration and explore how meaning imposes phenomenologically relevant constraints on information-based theories of consciousness.