The neocortex of higher mammals like carnivores and primates has a lattice-like network of local circuits embedded within inter-areal connections, which together form the ‘daisy architecture’ (DA). Our hypothesis is that the DA supports self-organized, context-dependent processing. Data derived from quantitative neuroanatomy and high spatio-temporal resolution imaging of cortical activity will be used to test theoretical predictions about the possible mechanisms and function of the DA. Understanding this architecture, and its computation will have two major benefits: Firstly, it will contribute to our understanding of how the brain reasons, and so will have important implications for mental health. Secondly, if the structure and dynamics of the DA could be reverse engineered it would be a major advance in Information Technology by offering novel methods for scalable, distributed, autonomous computation. We explore two candidate models of computation on the DA: graphical models such as Bayesian Networks and Factor Graphs that factor complicated global functions into a product of simpler ones and Dynamic Link Architectures that encode and recognize objects by dynamic composition of neuronal interactions. We implement these computational styles in hybrid analog/digital CMOS VLSI circuits, which contribute to the ‘End of Moore’s Law Problem’ by demonstrating how existing CMOS technology could be more efficiently deployed than in clocked digital systems. We explore the semantic nature of the biological computations by imaging the activity of cortical neurons as they respond to perceptually significant stimuli and evaluate our hypotheses about the DA by constructing object recognition systems that extract meaningful invariances from examples. These results would considerably advance our understanding of computation by the neocortex and provide a novel architecture in which implicit world semantics could be incorporated into computation.