Here's the learning path to master deep learning in 2020!
Introduction
Deep learning, a prominent topic in the domain of artificial intelligence, has been in the limelight for quite some time. He is especially known for his advancements in fields such as computer vision and gaming. (Alpha GO), exceeding human capacity. Since the last survey, there has been a drastic increase in trends. (click here to view the survey)
This is what Google trends show us:
If you are interested in the topic, here you have an excellent non-technical introduction. If you are interested in knowing the recent trends, here you have a great compilation.
Here we aim to provide a learning path for all those who are new to deep learning and also for those who wish to explore it further.. Then, Are you ready to embark on the journey of conquering deep learning? Let's go!
Paso 0: previous requirements
It is recommended that before jumping into deep learning, learn the basics of machine learning. The Machine Learning Learning Path is a comprehensive resource for getting started in the field.
If you want a shorter version, here it is:
Chronology : Suggested: 2-6 months
Paso 1: configure your machine
Before proceeding to the next step, make sure you have the compatible hardware. As usual, it is recommended that you have at least
- A good enough GPU (4+ GB), preferably Nvidia
- A CPU in good condition (as an example, Intel Core i3 is ok, Intel Pentium may not be)
- 4 GB RAM or depending on data set.
If you are still not sure, check this hardware guide.
PD: If you are a hardcore gamer (Not just candy grinders obviously!), It is possible that you already have the necessary hardware.
If you don't have the required specifications, you can buy it or lease a Amazon web service example. Here is a good guide to use AWS for deep learning.
Note: Do not install any deep learning libraries at this stage, do it in step 3.
Paso 2: a shallow dive
Now that you have a good enough understanding of the prerequisites, You must deepen your understanding of Deep Learning.
According to your preference, You can follow:
Along with the previous requirements, you must know the popular deep learning libraries and the languages to run them. Here is a list (Not complete) (see wiki page for more complete list):
Some other notable libraries include Mocha, neon, H2O, MXNet, Hard, Lasagna, Not learn. Here is a list of Deep Learning Libraries by Language.
To check Conference 12 by Stanford's CS231n curso for a brief overview of some of the popular libraries.
Chronology : 1-3 suggested weeks
Paso 3: Choose your own adventure!
Now comes the interesting part! Deep Learning has been applied in various fields with cutting edge results. To test this side of the moon, you, the reader, you can select which way to go. This should be a hands-on experience, so that you get a proper foundation on what you have understood so far.
Note: Each route contains a basic blog, a practical project, the deep learning library required for the project and an assistance course. Check the primer first, then install the required libraries and continue with the project. If you have any difficulties on the way, use the associated course to back it up.
- Deep learning for machine vision
- Deep learning for natural language processing
- Deep learning for voice / audio
- Deep learning for reinforcement learning
Chronology : 1-2 suggested months
Paso 4: deepening deep learning
Now this (almost) set to make a dent in the Deep Learning Hall of Fame! The road ahead is long and deep (pun) and mostly unexplored. It is now up to you to make use of this newly acquired skill as efficiently as possible.. Here are some tips to follow to hone your skill.
Chronology : Suggested – Infinity!
Noteworthy resources
Final notes
I hope this learning path has been useful to you. I have tried to make it as complete as possible. Now is the time for you to practice and read as much as you can. To gain experience working in neural networks, try our deep learning practice problem: Identify the digits.
Once you have understood deep learning and its associated concepts, do the Deep learning skill test. Familiarizing yourself with how deep learning is gaining accreditation is essential.
Good luck! Did you like reading this post? Do you follow an approach / package / different library to get started with Deep Learning? I would love to interact with you in the comments..
