analiza aiqbench
This paper proposes EGO, a novel software architecture designed to achieve Artificial General Intelligence through self-referentiality and autopoiesis, explicitly aiming to surpass the limitations of current Large Language Models. By grounding AGI in biological theories of cognition rather than standard neural network scaling, it represents a significant theoretical shift toward autonomous, embodied systems within AI research.
EGO: a Recursive and Self-Referential Cognitive Architecture for Artificial General Intelligence
arXiv:2507.14011v2 Announce Type: replace-cross Abstract: Artificial General Intelligence (AGI) is interpreted as an emergent property of autonomous and self-organizing systems, grounded in the principles of autopoiesis and embodied cognition, overcoming the structural limitations of current Large Language Models (LLMs). We introduce EGO (Environment Generative Operator), a software architecture based on the formal E-language, capable of self-referentiality and of maintaining its internal organization, thereby realizing Maturana's autopoiesis and providing a bridge between artificial intelligence and biological theories of cognition. For further details, please refer to the technical document on arXiv.
Abstract:Artificial General Intelligence (AGI) is interpreted as an emergent property of autonomous and self-organizing systems, grounded in the principles of autopoiesis and embodied cognition, overcoming the structural limitations of current Large Language Models (LLMs). We introduce EGO (Environment Generative Operator), a software architecture based on the formal E-language, capable of self-referentiality and of maintaining its internal organization, thereby realizing Maturana's autopoiesis and providing a bridge between artificial intelligence and biological theories of cognition. For further details, please refer to the technical document on arXiv.
Neural and Evolutionary Computing (cs.NE); Robotics (cs.RO)
arXiv:2507.14011 [cs.NE]
(or arXiv:2507.14011v2 [cs.NE] for this version)
https://doi.org/10.48550/arXiv.2507.14011 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 180 cuvinte