What would you recommend as the best resources for learning and teaching about Deep Learning? These could include tools (e.g. TensorFlow), tutorials, books (e.g. Deep Learning), MOOCs (e.g. Udacity's), etc.

I'm working to put together a list of best resources for an upcoming issue of AI Matters. What I'm looking for is the best of the best entry points into an understanding of deep learning. Nothing even so specific as convolutional networks is necessary. As an educator, I am motivated to collect and share the best first footholds for giving students and professionals a start for the mighty climb that is highly motivated by recent achievements. As far as audience, I envision undergraduate intro AI through adult professional.

  • $\begingroup$ Deep Learning Teaching Resources are Learning Teaching Resources which are Deep, and Learning Teaching Resources are Teaching Resources which are Learning. Not sure what you're asking about. $\endgroup$
    – AlMa1r
    Mar 18 at 1:06

4 Answers 4


The first thing that must be discussed in the matter, is the mathematics involved.

Deep learning is first and foremost a mathematical model. An overview of this big subject is given quite well in its Wikipedia article. The page showcases a summary of the subject, as well as links to more resources (2 birds in 1 go).

To explain the mathematical model, I recommend reading this article. This gives a relatively good basis of understanding for neural networks (which are a "supertype" of deep learning algorithm). Afterwards, page 4 of the article is a good place from which to progress towards Deep learning. For example, the neural network in the image from that page:

simple neural network

But what if we had more than 1 hidden layer? Now that's classified as deep learning.

A short google search for "deep learning neural networks" gives this handy website. It, too, gives a brief summary of the mathematics, increasing understanding.

After reading these articles and that website, you and your students should be good to go, and start gaining hands-on experience. Usually, the basic example for a deep learning algorithm is to fit a linear function to a dataset of points.

Another article (again, found through google search) explains the subject with more specific information. I recommend reading it to gain greater understand of the subject.

Because links sometimes go broken, it might be a good idea to download those articles (they are pdf files). Then again, they were found with straight forward google searches, so that's more reliable, especially in such a quickly evolving field.


From the perspective of MOOCs, a great place to start is MITx's 6.00.2x: Introduction to Computational Thinking and Data Science found on edX. It uses Python to introduce the study of data science and does not presume more than a beginner level of either Python or data science (although 6.00.1x, the first part of the course, is helpful for those who no experience whatsoever).

The penultimate unit focuses on Machine Learning using the following outline:

  • What is Machine Learning

  • K-Nearest Neighbors

  • Clustering

  • K-Means

For me it provided a successful introduction in two ways:

  • It defined machine learning with clear, simple examples

  • It made me want to learn more about the subject

For students who wish to study deep learning, a simple introduction to what machine learning even means and how it works at the simplest, most mundane of levels is a great jumping off point.

Additionally, CS50 at Harvard changed languages this year to include Python. As a result, they did a whole lecture on machine learning and deep learning in Python. A much more self-contained example, this lecture would certainly be another value starting point for beginning a dive into this discipline.

Links to CS50: Week 7 Lecture Video and Week 7 Lecture Notes

  • 1
    $\begingroup$ This would fit more into the category of general ML, but I appreciate your sharing the resource. Here are my recommendations for similar resources: https://sigai.acm.org/static/aimatters/3-2/AIMatters-3-2-05-Neller.pdf Thanks! $\endgroup$
    – ProfPlum
    Jun 12, 2017 at 18:38
  • $\begingroup$ @ProfPlum Agreed - I was just thinking that depending on the level of expertise, a short crash course on ML would be a good thing to have before diving into DL given no prior background. $\endgroup$
    – Peter
    Jun 12, 2017 at 21:14

This is only a partial answer as the scope of deep learning is vast and usually requires some fundamentals/prerequisites.

Google Developers Youtube channel has a playlist Machine Learning Recipes with Josh Gordon which while not definitive of deep learning, provides a very accessible introduction to the field, especially to a younger audience without getting too deep/technical into the mathematics. It uses Python (albeit version 2) and popular high level libraries as sci-learn and transiting to TensorFlow.


https://github.com/ChristosChristofidis/awesome-deep-learning has a long and curated list of resources. It should cover all possible ages and skill levels.

  • 3
    $\begingroup$ Welcome to CSEducators. While your link is probably helpful, can you say some more about what can be found there and why you think it will be helpful. We value longer answers more than such short ones, so that visitors can learn more before navigating away. Short answers are more likely to be ignored, but thanks for contributing. $\endgroup$
    – Buffy
    Jan 8, 2018 at 0:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.