What course do you recommend to learn about Deep Learning?
I want to learn more about Deep Learning, as in machine learning with neural networks. Is there a course that would give me practical knowledge in this area? #computer-science
4 answers
Joseph Catanzarite
Joseph’s Answer
Hi John. It's great that you want to learn this subject! Take Andrew Ng's 5-course Deep Learning Specialization (deeplearning.ai) on Coursera. This is by far the best introduction to deep learning I know of; it is not free but is well worth the moderate cost of $50 per month for ~6 months. You will learn neural networks from the bottom up, and you will gain broad understanding of fundamental concepts including how to set up a deep learning project, how to tune hyperparameters, how to improve your model, etc. You will be introduced to state-of-the art techniques in many of the main areas of deep learning, such as Convolutional Neural Nets, Object Detection, Object Location, Recommendation Engines, Face Recognition, Sequence Modeling, Natural Language Processing, Machine Translation. After this, take the (free) course "Practical Deep Learning for Coders" at fast.ai Best of luck on your deep learning journey!
Hi, John. I don't have any experience with Andrew Ng's course, but if you think of yourself more as a math/science person than a hacker, I would recommend http://neuralnetworksanddeeplearning.com/. This is how I learned about deep learning and I found that after going through this online book, I was ready to read papers in deep learning, learn a deep learning library, and start doing my own stuff. The book is a bit dated now, but I don't think this is really much of a problem except that in parts he uses Theano, which is a dying deep learning library probably not worth your time to learn. I would just skip those parts and instead learn PyTorch, which is a better library in every way.
Grant Reaber
The deeplearning.ai specialization was excellent.
Chris Curran
subramanian krishnamurthy
subramanian’s Answer
Python is a must.
Dinesh’s Answer
You need to have some knowledge of the following mathematical concepts:
- Probability and statistics
- Linear algebra
- Optimization
- Multivariable calculus
- Functional analysis (not essential)
- First-order logic (not essential)
Machine learning is about creating dynamic algorithms, which means your programming and software development skills will be put to test. This is much different from scripting web pages and creating simple windows applications. You’ll have lots to do with the fundamentals of analysis and design.
Dinesh recommends the following next steps:
Dylan’s Answer
I have taken the Coursera course Joseph listed and it was extremely helpful. The price $50 price tag is well worth! Coursera has a 2 week trail for every course so if you are on the edge about dropping $50 you can test the waters and see if the course is for you!
But Coursera has many other courses on Deep Learning! Don't feel confined to just this one! Test out the others and see which learning program best fits your needs!
Udacity (https://www.udacity.com) also has many courses on deep learning! Some are free and some are paid but these are also well worth the time! I personally started my deep learning journey through a course on Udacity and it was a very helpful and soft introduction.
https://classroom.udacity.com/courses/ud188
This is a free introduction to deep learning using Pytorch offered by Udacity. This course will go over the math and fundamentals behind deep learning as well as the applications of a deep learning model!