Why mathematics is important for machine learning?
Answer depends who you ask this question – “Why math is important in machine learning?”. Because there are so many roles using data science and machine learning. Managers, developers, analysts, data scientists working in different fields and for different purpose. Some are focused on getting insights out of data. Some are focused on telling stories from data. Some are using tools for most common tasks like regression, classification, clustering. Some are building intelligent products. Some are doing research. Some are doing innovation.
The truth is there are places where you are nothing without math, and there are places where your math is just a joke.
Although there are many plug and play libraries and packages like sklearn, tensorflow, weka and many more. Having mathematical intuition is important because machine learning does not start and end with just tools.
Here is why mathematics is so important for machine learning:
To Build The Intelligence.
The number one reason why you need mathematics is: “to build the intelligence”.
To build the intelligence you need mathematics, There is no other way around it. Sometimes you have to design a new algorithm, you have a problem to solve and you have to derive the mathematical formula to solve that problem.
All the machine learning problems are not just classification, regression and clustering. There is more. When your goal is to build intelligence into the system, tools can be just a small part of it, not everything. So it is very important to understand Linear Algebra, Multivariable Calculus and Optimization, Probability & Statistics in the context of machine learning.
Machine Learning is a research and Innovation field.
Number two reason is “research and innovation”. There is math behind every new research and innovation.
Sure you can be a data scientist and keep running scikit-learn and tensorflow your whole life, but remember machine learning is a field of research and innovation. There is nothing much you can achieve with existing tools and techniques. Mathematics is one of the most important skill required to build something new in machine learning.
Understanding the logic behind algorithms so that you can choose best and fine tune.
Now most people in industry don’t talk about math, because most people don’t use it. And even if they use it’s just limited to their understanding of how machine learning algorithms work, and that requires at-least some basic understanding of math. so that they can fine tune the model for their purpose.
If you passionate about the data science field and possibilities math will be definitely helpful.
💡 Machine learning is a math poetry and you can’t enjoy a poetry without having a good understanding of it’s language.
💡 After you have some experience with machine learning, it’s good to look back at fundamental mathematics and you will see what you have missed first time.
All the best.