SDS 474: The Machine Learning House — with Jon Krohn

Core Subjects

Jon Krohn emphasizes the importance of foundational subjects like linear algebra, calculus, and probability in the field of data science. These subjects form the bedrock of statistical models and machine learning approaches, remaining constant despite the ever-evolving tools and techniques in the industry 1. Jon suggests that while high-level tools like scikit-learn and Keras are exciting, a deep understanding of these core subjects is crucial for long-term success.

The foundational subjects that underlie data science, including those behind statistical models and machine learning approaches, that is, linear algebra, calculus, probability theory, and data structures. These subjects barely change at all, decade after decade after decade.

--- Jon Krohn

He believes that mastering these areas not only enhances one's ability to innovate but also facilitates the transition to specialized domains like deep learning and natural language processing 1.

   

Application

Understanding foundational knowledge is essential for applying machine learning in real-world scenarios. Jon explains that these subjects enable data scientists to improve models and deployments by providing a deeper appreciation of machine learning theory 1. This understanding is likened to the ground floor of a 'machine learning house,' supporting more advanced applications.

To be an outstanding data scientist or machine learning engineer, it doesn't suffice to only know how to use models via the abstract interfaces that the most popular libraries provide.

--- Jon Krohn

Jon highlights that a strong grasp of these fundamentals is crucial for transitioning from general machine learning principles to specialized areas, as these often require knowledge found in academic papers or textbooks 1.