Skip to main navigation menu Skip to main content Skip to site footer
##common.pageHeaderLogo.altText##

Philosophy and Political science

December 22, 2023; Boston, USA: V International Scientific and Practical Conference «SCIENTIFIC PRACTICE: MODERN AND CLASSICAL RESEARCH METHODS»


COMPARISON OF HUMAN AND MODERN LARGE LANGUAGE MODELS THINKING


DOI
https://doi.org/10.36074/logos-22.12.2023.070
Published
04.01.2024

Abstract

Thinking is a complex process, which humans try to understand since the very beginning of our existence. It includes generation of thoughts, making sense of the existing experience, analysis, drawing conclusions and making conclusions. Modern artificial intelligence systems are trying to replicate this process for different domains starting with texts and now even audio or video. However, generation of new knowledge, correct answers from existing knowledge or solving complex tasks requires step-by-step approach, analysis, sources comparison and even full-on planning or ability to reassess previous decisions. We try to describe the thinking process in this article, derive main properties and compare it to the way modern artificial intelligence models learn information and make decisions. Nowadays the necessity to understand this process more clearly is as relevant as ever, because it will give us an opportunity to classify current or future models as “artificial general intelligence” [1].

References

  1. Chollet, F. (2019). On the measure of intelligence. arXiv preprint arXiv:1911.01547.
  2. LEIGH F. Platonic Dialogue, Maieutic Method and Critical Thinking. Journal of Philosophy of Education. 2008. Vol. 41, no. 3. P. 309–323. URL: https://doi.org/10.1111/j.1467-9752.2007.00561.x (date of access: 18.12.2023).
  3. Kant I. I. Immanuel Kant Critique of Pure Reason:(Annotated Edition). Independently Published, 2021.
  4. Kahneman D. Thinking, Fast and Slow. Worth Publishers, Incorporated, 2013.
  5. Chang E. Y. Prompting Large Language Models With the Socratic Method. 2023 IEEE 13th Annual Computing and Communication Workshop and Conference (CCWC), Las Vegas, NV, USA, 8–11 March 2023. 2023. URL: https://doi.org/10.1109/ccwc57344.2023.10099179 (date of access: 18.12.2023).
  6. Schrittwieser, J., Antonoglou, I., Hubert, T., Simonyan, K., Sifre, L., Schmitt, S., ... & Silver, D. (2020). Mastering atari, go, chess and shogi by planning with a learned model. Nature, 588(7839), 604-609.

Most read articles by the same author(s)