Loading...

Artificial Neural Network Functionality Analysis using Mathematical Logic: A Survey

Hadizadeh Moghadam, Ali | 2025

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 58683 (02)
  4. University: Sharif University of Technology
  5. Department: Mathematical Sciences
  6. Advisor(s): Ardeshir, Mohammad
  7. Abstract:
  8. Neural Networks, and its many variants, are currently one of the main technologies in the domain of Artificial Intelligence, and they are receiving more attention than ever due to the increasing amount of responsibilities they are entrusted with. These include a myriad of tasks in medicine, engineering, and entertainment, to name a few; and involve skills such as interacting with humans, pattern recognition in images, and editing and creating textual or visual content. However, in spite of the steady rise of human reliance on AI, there are yet many questions to be answered regarding its functionality. For instance, we still cannot deduce precisely how an AI agent arrives at its answers, nor can we predict how much resource does it need to achieve an acceptable performance. Two approaches come to mind: to inspect and interpret each agent one by one, or to investigate the theoretical strengths and constraints shared by all agents of a kind. The thesis at hand concerns itself with the theory of Mathematical Logic and the toolset it provides us as we examine the aforementioned questions. Making use of this theory could be justified by the fundamental connections between mathematical logic, computation, and the human mind. This research aims to provide a critical survey of previous attempts and achievements in this regard
  9. Keywords:
  10. Artificial Intelligence ; Neural Networks ; Mathematical Logic ; Artificial Neural Network

 Digital Object List

 Bookmark

No TOC