I'm a biotech student and I'm currently working on single-particle tracking. For my work, I need to use aspects of deep learning (CNN, RNN and object segmentation) but I'm not familiar with these topics. I have some prior knowledge in python.
So, do I have to learn machine learning first before going into deep learning, or can I skip ML?
What are the pros and cons of studying machine learning before deep learning?
That question doesn't really make sense: deep learning is a sub-topic of machine learning, so you can't really 'skip' it. It's a bit like "I want to learn about trigonometry, but do I need to do geometry first?"
Having said that, in order to make sense of deep learning you should really know about the general principles of machine learning, otherwise you won't understand it. Or, more importantly, you won't understand what problems deep learning can be applied to, and what issues are better solved with other methods.
You don't need to go into much detail, but should at least get an overview.
Like Oliver Mason mentioned, Deep learning is just a sub-field of machine learning. In order to learn deep learning effectively you need to have certain pre-requisites like basic principle of Machine learning and basics of simple Artificial neural network with some programming knowledge ( Python is go-to language). That being said, you don't need to know every single Machine learning algorithm and it's practices.
Now if deep learning happens to be just a tool that you need for this particular project and have no time to learn in depth about it then I would recommend you to take a look at python libraries like Tensorflow, pytorch, scikit learn, scipy, open cv etc. You can get started and use DL, ML models with these and many other libraries without knowing it's under the hood algorithms and implementations.
One of the best course to get started with deep learning with very little Ml knowledge is Andrew ng's deep learning.ai course on coursera ( you can audit the course and get all the course materials for free)
Here's the link to the course : Deep learning.ai
Machine learning uses algorithms to digest data sets, draw conclusions based on analyzed data, and use these conclusions to complete the task in the most effective way. This ability is a fundamental difference between machine learning and machine that has been programmed from the beginning with a certain sequence of commands. Machine learning has the capability to accomplish tasks dynamically.
While Deep Learning is one of the methods of implementing machine learning that aims to mimic the workings of the human brain using ANN. Deep learning uses a number of algorithms as 'neurons' to work together in determining and digesting certain characteristics in a data set.
In contrast to general machine learning programs that are designed to accomplish certain tasks, deep learning programs are usually programmed with more complex capabilities to study, digest, and classify data.
A machine learning model requires data to learn and obtain parameter estimates, so the more data that can be used, the machine learning program will be smarter. In addition, operating machine learning models — especially logical networks for deep learning — requires high computational power. This is because the deep learning model must operate many processes simultaneously, especially in the training phase. In the training phase, the machine learning model must process very large amounts of data to be categorized as a reference.
So, do I have to learn machine learning first before going into deep learning, or can I skip ML?
Quoting from Wikipedia, "Deep learning (also known as deep structured learning or differential programming) is part of a broader family of machine learning methods based on artificial neural networks with representation learning."
That being said, it is better to understand the fundamental of machine learning first so you would be able to understand completely of how deep learning works, and how to apply it effectively and efficiently. However, You can always skip straight to deep learning without having any major issues, as there is already a lot of libraries supporting deep learning on python such as TensorFlow.
What are the pros and cons of studying machine learning before deep learning?
pros:
Since deep learning is a subset of machine learning, having fundamental knowledge about machine learning and the other machine learning algorithms will be beneficial.
cons:
You might (not) waste your time and energy to learn something you wouldn't use.
