New Theory Explains Intelligence as Emergent Phenomenon
Does intelligence amount essentially to the learning and imitation of examples, as the current paradigm of machine learning suggests, or is it dominated by structure that is given a priori? In a new manuscript, a team of ZHAW and ETH/UZH presents convincing arguments for the latter.
To this day the human and animal brain with its incredible ability to pursue vital goals in complex natural environments defies understanding.
In the as yet unpublished article "A Theory of Natural Intelligence" (Lead author Christoph von der Malsburg; siehe https://arxiv.org/pdf/2205.00002.pdf) a team of the ZHAW's CAI Group Computer Vision, Perception and Cognition, ETH/UZH's INI and the Frankfurt Institute for Advanced Studies proposes a totally novel explanation for this phenomenon. According to it the structure of the brain arises by emergence and is characterized by strong regularity that mirrors, in a natural way and ahead of all learning, the basic structure of the environment. Implicit in this structure is the capacity to enact genetically encoded abstract behavioral schemata in concrete situations.
Besides putting Epistemology on a new grounding, the paper lays the basis for the artificial emulation of animal and human intelligence, and with that for a technology of autonomous vehicles and robots. First applications of the new theory in the context of machine vision are currently attempted in a master thesis project.