The instructor describes: Improvement methods are often used to tune hyperparameters in machine learning through iterative training of the model using mathematical function evaluation. The improvement process helps the model deliver optimal performance. This course provides an in-depth overview of modern mathematical improvement methods applied in machine learning and data science. The course covers topics: Introduction to curved improvement models in data science, curvature tools and the calculation of non-smooth differential calculus for improvement, subgradient methods, proximal gradient methods, accelerated gradient methods (momentum), limits and errors of learning, and an introduction to non-curved improvement models in supervised machine learning.
Credit hours: 3
Objectives of the course :
• Identifying the problems and challenges arising in machine learning and the modern analysis of data such as complexity, robustness, and implementability.
• Understanding the basics of mathematical optimization for machine learning.
• Understanding the various methods of improvement that support today's popular machine learning models in real-world applications.
• Participation in designing methods for improving machine learning and learning from practice and experience.
Course outputs :
1. Knowledge and understanding:
1.1. Identifying categories and improvement issues in machine learning.
1.2. Understanding the basic concepts of improvement and advanced improvement algorithms.
2. Skills:
2.1. Classification of improvement issues based on their feasibility, solution, difficulty, and compatibility with existing software.
2.2. Applied participation in designing methods to improve machine learning.
3. Values:
3.1. Solving a collective problem related to methods of improving machine learning.
3.2. Applying improvement solutions in an ethical manner in scientific research.
This website uses special cookies to ensure ease of use, improve your browsing experience, and clarify the terms and policies related to About user privacy. By continuing to browse this website, you acknowledge that you accept the use of cookies and the terms of the Privacy Policy