终极智能第二部分：智能的物理度量与复杂度
人工智能
2016-05-12 v2
摘要
我们继续分析适用于量化归纳推理系统的体积和能量度量。我们将算法信息论（AIT）中的逻辑深度和概念跳跃大小度量扩展到随机问题，以及涉及体积和能量的物理度量。我们引入了一种我们认为适用于智能机器的计算复杂度图形模型。我们展示了归纳推理中能量、逻辑深度和计算体积之间的几个渐近关系。特别是，我们得出了归纳推理的“黑洞方程”，该方程将最优归纳推理解的能量、体积、空间和算法信息联系起来。我们引入了能量有界算法熵。我们简要地将我们的思想应用于宇宙中智能计算的物理极限。
引用
@article{arxiv.1504.03303,
title = {Ultimate Intelligence Part II: Physical Measure and Complexity of Intelligence},
author = {Eray Özkural},
journal= {arXiv preprint arXiv:1504.03303},
year = {2016}
}
备注
This paper was initially submitted to ALT-2014. We are taking the valuable opinions of the anonymous reviewers into account. Many thanks to Laurent Orseau for his constructive comments on the draft, which inspired this revision. arXiv admin note: substantial text overlap with arXiv:1501.00601 This is a major revision over the last version edited