EUVIMEDEUVIMED entdecken
Crossref

On Measurable Tournaments for Progressing Generalized Cognizers

Edward Pogossian

Mathematical Problems of Computer Science · 01.06.2026 · 65 · 49-62

DOI: 10.51408/1963-0145 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Mathematical Problems of Computer Science, online erschienen am 01.06.2026. Verfasst von Edward Pogossian.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Researchers such as Kurzweil and Goertzel predict that AI, due to the progress in LLM is entering a period of exponential growth toward Artificial General Intelligence (AGI). They believe that if such AGI is capable of rewriting its own code, it could evolve into a superhuman AI, possessing the cognitive and computational power of all human civilization We interpret this requirement for AGI as an unavoidable ability of generalized cognizers to generate expectably reasonable versions of development of themselves, along with trustworthy means of measuring these versions and selecting the most promising of them wrt the utilities of cognizers. Generalizing our prior assertion that knowledge-based chess strategies can be strongly scaled by local tournaments, we argue an analogous statement for suitable measuring of the progress of generalized cognizers in the frame of an adequate theory of cognizing and address to the ways to strengthen it.

Zitieren

APA

Pogossian, E. (2026). On Measurable Tournaments for Progressing Generalized Cognizers. Mathematical Problems of Computer Science, 65, 49-62. https://doi.org/10.51408/1963-0145

Vancouver

Pogossian E. On Measurable Tournaments for Progressing Generalized Cognizers. Mathematical Problems of Computer Science. 2026;65:49-62. doi:10.51408/1963-0145

BibTeX
@article{Pogossian2026,
  author = {Pogossian, Edward},
  title = {On Measurable Tournaments for Progressing Generalized Cognizers},
  journal = {Mathematical Problems of Computer Science},
  year = {2026},
  volume = {65},
  pages = {49--62},
  doi = {10.51408/1963-0145}
}
Kostenlos starten →