On the Definition of Intelligence
This work addresses the foundational challenge of defining intelligence for AGI engineering, but it is incremental as it builds on existing concepts without presenting new empirical results.
The paper tackles the problem of defining intelligence for AGI by proposing a general criterion based on entity fidelity, where intelligence is the ability to generate entities exemplifying a concept given examples, formalized as ε-concept intelligence with a tolerance ε.
To engineer AGI, we should first capture the essence of intelligence in a species-agnostic form that can be evaluated, while being sufficiently general to encompass diverse paradigms of intelligent behavior, including reinforcement learning, generative models, classification, analogical reasoning, and goal-directed decision-making. We propose a general criterion based on \textit{entity fidelity}: Intelligence is the ability, given entities exemplifying a concept, to generate entities exemplifying the same concept. We formalise this intuition as \(\varepsilon\)-concept intelligence: it is \(\varepsilon\)-intelligent with respect to a concept if no chosen admissible distinguisher can separate generated entities from original entities beyond tolerance \(\varepsilon\). We present the formal framework, outline empirical protocols, and discuss implications for evaluation, safety, and generalization.