Deep Learning
Deep learning is a subset of machine learning, which is essentially a neural network with three or more layers. These neural networks attempt to simulate the behavior of the human brain—albeit far from matching its ability—allowing it to “learn” from large amounts of data. While a neural network with a single layer can still make approximate predictions, additional hidden layers can help to optimize and refine for accuracy.
Neural Networks
This is where everything changed. Neural Networks are the most beautiful and accurate mathematical representation of a biological neuron. Well not the most accurate but still close, quite close.
Here's what Sam Altman, the founder of OpenAI says about NNs:
Neural networks really, truly learn. It's not a fancy trick. This is one of the most remarkable things humans have ever figured out, and the implications are difficult to overstate.
So, now you know that something called Neufvwel Net... bluh bluh exists. Well what good is it, how can it help achieve that which we might have missed all along. To know that you gotta learn how NNs work:
Once again, 3B1B to the rescue. I highly recommend checking out the Neural Networks playlist of his. Really opens up a world of possiblities.
Make Your Own Neural Network - Tariq Rashid: This book is really interactive and helps understand the fundamentals quite easily. It's in epub format so make sure you have some software to run that file.
Neural Networks and Deep Learning - Michael Nielsen: It's a free online book and probably one of the best if it comes to Neural Networks and their concepts. Really good stuff.
Deep Learning - Ian Goodfellow: I do not recommend this book for beginners. While it is state of the art for DL, but it's got too much math in it, which might be a little overwhelming for first time learners. Keep it in your box for later though. Or feel free to explore.
Deep Learning Fundamentals
Once you get the understanding of how a neural network works, we can move forward to deep learning.
| Course | Format | Details |
|---|---|---|
| Kaggle's Intro to Deep Learning | docs | Really good place to get started with the concepts. |
| Andrew Ng | videos | Andrew again, he is the guy I'm telling you. This course gives a full fledged understanding of a lot of DL concepts. Take your time with this. Here you can find the notes of this course. |
| Deep Learning: State of the Art | videos | This is an interactive introduction to deep learning by Lex Fridman |
| Intro To Deep Learning: MIT | videos | This consists of everything from Reinforcement Learning, RNN, GAN etc. |
| Convolutional Neural Networks for Visual Recognition: Stanford University | videos | High quality material from Stanford University on CNNs and visual image recognition. |
| Practical Deep Learning for Coders | docs/videos | This is hands on implementation of Deep Learning using fastai and Pytorch. Good if you have some experience in the field. |
| Deep Learning with Python | book | This book on Deep Learning is beginner friendly and delves into it's implementation using Python,Keras and TF. |
Mathematical Deep Learning
Alert: This is not for everyone. Don't even look at it if you're getting started.
These are helpful for
Research Enthusiasts.
Below are the links of lectures from Prof. Ali Ghodsi - University of Waterloo. His explanations are best for understanding deep learning and the mathematics behind it. Refer to this in case you want to go deep, really deep into the mathematics behind deep learning.
| Course | Format | Details |
|---|---|---|
| Introduction: Part 1 | videos | Part 2, Part 3 |
| Backpropagation | videos | Math behind backpropagation algorithm |
| Regularization: Part 1 | videos | Part 2, Part 3 |
| PCA/ LDA/ QDA/ Decision Tree: Part 1 | videos | Part 2, Part 3 |
| Word2Vec: Part 1 | videos | Part 2, Part 3, Part 4 |
| Tree methods and Boosting: Part 1 | videos | Part 2, Part 3 |
| SVM: Part 1 | videos | Part 2, Part 3 |
| RNN: Part 1 | videos | Part 2, Part 3 |
| CNN: Part 1 | videos | Part 2, Part 3 |
Woosh!! Well, that was deep (no pun intended 😵💫)
I guess this much stuff will give you a pretty good idea of deep learning. Feel free to explore by yourself too and tell us if you find something interesting. We'll add it here.
Ciao 👋🏻.