Work

Secret Mathematics of Agnosticism

Year: 2026

Type: preprint

Abstract: Agnosticism usually is classified as an informal philosophical viewpoint or hesitation on theological aspects. However, this paper presents agnosticism as a mathematically optimal state of ignorance represented by the maximal entropy, vacuous probability intervals in the context of Dempster-Shafer theory, and an invariant fixed point during Bayesian belief update. The framework of the empty/epistemic simplex and corresponding entropic dynamics is introduced. Moreover, problems of imprecise probability in the case of radical ignorance and algebraic multi-valued logic are discussed. Finally, open problems and research avenues are suggested for the implementation of mathematical agnosticism in Artificial General Intelligence (AGI) and distributed knowledge management systems.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.23891820

Author: Mageed, Ismail A

Language: English

Topic: Bayesian Modeling and Causal Inference

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7225080159

Cites:

Cited by:

Related to: 10