通用认知智能:从认知意识到Lambda(Λ)
1. 引言与文章计划
通用认知智能(UCI)的概念可在很大程度上从先前引入并公理化的认知意识理论(TCC)以及用于衡量给定时间(或一段时间间隔)内智能体认知意识程度的框架 Λ 中推导得出。我们不会简单重复之前的内容,而是在此基础上引入适用于所有认知智能体的 UCI,并介绍如何轻松构建 Λ 的新的无穷版本。
为了初步刻画 UCI,我们先通过一个测试来探讨认知智能。假设数学老师 Alice 提出一个问题:已知命题 ¬(ϕ → ψ),是否能演绎得出 ϕ 成立?并要求证明答案的正确性。答案是肯定的,以下是一个详细的证明过程:
Proof: Our given is that ¬(ϕ →ψ).
Suppose for indirect proof
that ¬ϕ holds. Suppose in addition, for a sub-proof using reductio
as well, that ¬ψ; and suppose too — to obtain a conditional directly
inconsistent with our given — that ϕ. We deduce ψ by reductio from
the now-available contradiction, and then since our supposition of ϕ
has led to ψ, we deduce the conditional ϕ →ψ. But now we have
before us the contradiction of this conditional with our given, a
超级会员免费看
订阅专栏 解锁全文
26

被折叠的 条评论
为什么被折叠?



