A General Intelligence Theory Based on the Representation of Need
Bowen Sun, Xiaojun Mei, Yi Du, Mengqi Gao · 2024
Proposed at the Dartmouth Conference in 1956, the primary objective of artificial intelligence (AI) is to empower machines to think and solve problems akin to humans. Despite the emergence of various large-scale models in recent years, notable progress has been made in the realm of artificial general intelligence (AGI) models. However, machines still encounter challenges in autonomously learning and evolving like humans, and there remains a dearth of a robust mathematical foundation. This paper delves into the autonomous evolutionary process observed in humans and biological entities, asserting that “Needs are the fundamental driving force for organism evolution.” Additionally, the paper introduces corresponding representation algorithms and lays out mathematical theoretical foundations, thereby furnishing a theoretical explanation and a solid mathematical basis for understanding the autonomy of general intelligence models.