arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2205.14718cs.CCcs.ITmath.IT
收藏

概率Kolmogorov复杂性理论及其应用

Theory and Applications of Probabilistic Kolmogorov Complexity

Zhenjian Lu, Igor C. Oliveira

摘要:

近年来,Kolmogorov复杂性在学习[CIKK 16],电路复杂性[OPS 19],密码学[LP 20],平均情况复杂性[Hir21]和证明搜索[Kra22]中的各种应用已经被发现。由于算法的运行时间是这些领域的关键资源,因此在相应的参数中考虑Kolmogorov复杂性的时间有界变量是至关重要的。虽然有时间限制的柯尔莫哥洛夫复杂性和理论计算机科学的不同领域之间富有成效的相互作用已经知道了相当长的一段时间(例如,[Sip 83,Ko91,ABK+06,AF 09],仅举几例),上述结果导致了对该主题的新兴趣。Kolmogorov复杂性的理论是很好理解的,但许多有用的结果和性质的Kolmogorov复杂性并不知道在时间有界的设置。这就造成了技术上的困难,或导致条件的结果时,应用方法从时间有界的柯尔莫哥洛夫复杂性的算法和复杂性理论。也许更重要的是,在许多情况下,有必要考虑随机算法。由于随机字符串具有很高的复杂性,经典的时间有界柯尔莫哥洛夫复杂性理论在这种情况下可能不合适。为了缓解这些问题并发展一种在随机计算环境中生存的时间有界柯尔莫哥洛夫复杂性理论,最近的一些论文[Oli19,LO21,LOS21,GKLO22,LOZ22]探索了时间有界柯尔莫哥洛夫复杂性的概率概念,例如$\mathsf{rKt}$复杂性[Oli19],$\mathsf{rK}^t$复杂性[LOS21]和$\mathsf{pK}^t$复杂性[GKLO22]。这些措施考虑通过概率表示对对象进行编码的不同方式。在本次调查中,我们介绍了概率时间有界Kolmogorov复杂性及其应用,重点介绍了许多悬而未决的问题和研究方向。

英文摘要:

Diverse applications of Kolmogorov complexity to learning [CIKK16], circuit complexity [OPS19], cryptography [LP20], average-case complexity [Hir21], and proof search [Kra22] have been discovered in recent years. Since the running time of algorithms is a key resource in these fields, it is crucial in the corresponding arguments to consider time-bounded variants of Kolmogorov complexity. While fruitful interactions between time-bounded Kolmogorov complexity and different areas of theoretical computer science have been known for quite a while (e.g., [Sip83, Ko91, ABK+06, AF09], to name a few), the aforementioned results have led to a renewed interest in this topic. The theory of Kolmogorov complexity is well understood, but many useful results and properties of Kolmogorov complexity are not known to hold in time-bounded settings. This creates technical difficulties or leads to conditional results when applying methods from time-bounded Kolmogorov complexity to algorithms and complexity theory. Perhaps even more importantly, in many cases it is necessary to consider randomised algorithms. Since random strings have high complexity, the classical theory of time-bounded Kolmogorov complexity might be inappropriate in such contexts. To mitigate these issues and develop a theory of time-bounded Kolmogorov complexity that survives in the setting of randomised computations, some recent papers [Oli19, LO21, LOS21, GKLO22, LOZ22] have explored probabilistic notions of time-bounded Kolmogorov complexity, such as $\mathsf{rKt}$ complexity [Oli19], $\mathsf{rK}^t$ complexity [LOS21], and $\mathsf{pK}^t$ complexity [GKLO22]. These measures consider different ways of encoding an object via a probabilistic representation. In this survey, we provide an introduction to probabilistic time-bounded Kolmogorov complexity and its applications, highlighting many open problems and research directions.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑