Computation, Condensation, and the Incompleteness Between Them: A Coupled Foundation of Intelligence
arXiv:2303.04203v4 Announce Type: replace Abstract: The theory of computation was built to answer Turing's question: what is effectively calculable by an unbounded, immortal, disembodied agent following rules? Intelligence answers a different question (nature's): what can a \emph{finite}, mortal, energy-limited agent do quickly enough to survive in a non-stationary world? We argue that a complete answer requires two operators: \emph{computation} and \emph{memorizaion}. Computation, $\dpar$, transforms structure toward closure; memorization, $\kap$, condenses a validated closed cycle into a reu
The explosion of AI capabilities and the increasing focus on embodied AI and practical rather than theoretical intelligence necessitate a re-evaluation of fundamental definitions.
A refined theoretical foundation for intelligence, integrating both computation and memorization, could guide future AI research and development towards more robust and adaptive systems.
The proposed framework diverges from purely computational views of AI, suggesting that practical intelligence in finite agents requires an interwoven process of transformation and consolidation.
- · AI researchers focusing on embodied cognition
- · Developers of memory-augmented AI architectures
- · Philosophers of AI
- · Strictly theoretical computational AI paradigms
- · AI approaches ignoring memory's role in intelligence
New AI models emerge that explicitly integrate computation and memorization, leading to more efficient learning.
This foundational shift influences educational curricula and research priorities in computer science and AI.
More generalizable and adaptable AI agents accelerate progress in areas like robotics and complex decision-making, potentially altering labor markets more quickly.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG