Loading...

Implementation and Comparison of Some Optimization Methods from the Perspective of Machine Learning

Tabesh, Aryan | 2024

0 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 57707 (02)
  4. University: Sharif University of Technology
  5. Department: Mathematical Sciences
  6. Advisor(s): Mahdavi Amiri, Nezamoddin
  7. Abstract:
  8. Machine learning is advancing remarkably, bringing profound theoretical insights and finding applications in various fields. Optimization is a crucial area in machine learning, attracting the attention of many researchers. As the volume of data increases and models become more complex, the challenges faced by optimization methods are becoming more diverse and complicated. Continuous efforts are being made to overcome these challenges through new solutions and improvements to existing methods. Understanding and summarizing these optimization methods from a machine learning perspective is beneficial and essential, leading to further advancements in both fields. Here, based on the collected materials available in the article and the literature, we first examine the various optimization challenges in machine learning. Then, we delve into the fundamental principles and recent advancements in common optimization methods. Following that, we demonstrate how these methods are adapted for application in several key areas of machine learning. Subsequently, we discuss the unresolved challenges and questions in the field of optimization, present perspectives and directions for future research, and finally, in this thesis, we have implemented various algorithms and tested the developed programs
  9. Keywords:
  10. Machine Learning ; Optimization Method ; Deep Learning ; Large Scale Machine Learning ; Big Data

 Digital Object List

 Bookmark

No TOC