Saved in:
Bibliographic Details
Main Author: Ng, Kei-Sing
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.22423
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913990389530624
author Ng, Kei-Sing
author_facet Ng, Kei-Sing
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2507_22423
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle On the Definition of Intelligence
Ng, Kei-Sing
Artificial Intelligence
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.
title On the Definition of Intelligence
topic Artificial Intelligence
url https://arxiv.org/abs/2507.22423