A formal mathematical theory of intelligence as the ability to achieve goals across a wide range of environments — measured against the universal algorithmic prior.
AIXI (Hutter): a formal definition of the optimal agent — an agent that maximizes expected reward across all computable environments, using Solomonoff's universal prior. Universal intelligence (Legg & Hutter): intelligence = the ability to achieve goals in a wide range of environments. Formally: the expected reward of an agent over all possible computable environments, weighted by their Kolmogorov complexity. Provides a mathematical foundation for comparing agents by intelligence. Practically uncomputable but theoretically foundational.
Discussing
Ask anything about Universal Intelligence Theory — their ideas, influences, contradictions, legacy.
Sign in to start a conversation.