Creativity, Learning, and Neuromorphic Design in Generative AI

Author

Alek Zyszkowski

Date of Award

2024

First Advisor

Anne O'Dwyer

Second Advisor

Jennifer Daniels

Abstract

This thesis aims to examine contemporary Artificial Intelligence (AI) models in light of the neurobiological and psychological factors that go into conscious and unconscious cognitive learning processes. Explored are how such ideas have previously been implemented into machine learning and how newer findings can potentially enrich new Artificial Intelligence projects such as “deep learning” and large language models. The analysis builds on the research conducted on AI and neurobiological learning functions and associated theories, using information and ideas from reinforcement learning, including long term potentiation and reward systems such as dopamine. In the latter portion of the project, I make inferences and identify possible applications /solutions in contemporary and future neurological and AI endeavors. Using my own experimentations along with the others' research, I hope to display the evident difference between current Generative AI and where it is still limited from a computational and psychological perspective. In conclusion, I explore how one may use these ideas to devise potential new computer projects, and create neurobiological and psychological understandings of AI.

Recommended Citation

Zyszkowski, Alek, "Creativity, Learning, and Neuromorphic Design in Generative AI" (2024). Senior Theses. 1687.
https://digitalcommons.bard.edu/sr-theses/1687