Can I start learning deep learning without taking a machine learning course?



I’m new to Machine Learning and Deep Learning.

I read online that you do not need to know Machine Learning for learning Deep Learning. Is it true? Please guide me on the same.



Hi Jim Wright,

If you are comfortable with the following concepts in machine learning you can take the deep learning course:

  • Ideas such as Instances, Labels, Objective Function
  • The concepts such as Gradient Descent
  • Underfitting / Overfitting

Sandeep Giri


Hi @Wright_Jim,

The short answer is Yes, though I do not recommend it.

If you understand Gradient Descent, classification, overfitting, regularization, cost function etc then you can learn Deep Learning without taking a Machine Learning course. Deep neural networks outperform traditional machine learning algorithms provided you have a good amount of data.

Though deep learning is one of the most exciting fields of machine learning, in my opinion, you should not jump to Deep Learning straight away. First master the fundamentals. Most of the problems can be solved well with techniques such as Random Forests and Ensemble. Deep Learning is best suited for the complex problems such as image recognition, speech recognition provided that you have the sufficiently large amount of data, computing power, and most important patience :). I’ve seen my neural network training goes on for hours even for days sometimes. So in case you do not have that complex problem or computing resources then I will recommend to use Machine Learning algorithms first.

At CloudxLab, we are building a platform to learn machine learning and deep learning in a gamified way so that you learn faster and better. This specializationcourse starts from the foundations, then deeps dive into the advanced concepts machine learning and deep learning.

Hope this answer helps you and all the best for deep learning course. Believe me, you will love coding your deep neural nets :slight_smile:

PS - I’ve written it originally on quora