A machine learning craftsmanship blog.
Video Walkthrough: Summary: DECENTRALIZED AI!!!!!!! OK let's do this. We're gonna build a P2P AI in about 50 lines of Python. Your friends text you on WhatsApp, a local AI responds using your local data, and sender-specific context folders protect privacy. If that's boring and you want the fancy stuff (homomorphic encryption, blockchain, federated learning, etc.), maybe still start here. This is…
TLDR: I've recently wondered about whether the Universe is truly random, and I thought I'd write down a few thoughts on the subject. As a heads up, this post is more about sharing a personal journey I'm on than teaching a skill or tool (unlike many other blogposts). Feel free to chat with me about these ideas on Twitter as I'm still working through them myself and I'd love to hear your…
TLDR: What if it was possible for surveillance to only invade the privacy of criminals or terrorists, leaving the innocent unsurveilled? This post proposes a way with a prototype in Python. Edit: If you're interested in training Encrypted Neural Networks, check out the PySyft Library at OpenMined Abstract: Modern criminals and terrorists hide amongst the patterns of innocent civilians, exactly…
TLDR: In this blogpost, we're going to prototype (from scratch) and learn the intuitions behind DeepMind's recently proposed Decoupled Neural Interfaces Using Synthetic Gradients paper. I typically tweet out new blogposts when they're complete at @iamtrask . Feel free to follow if you'd be interested in reading more in the future and thanks for all the feedback! Part 1: Synthetic Gradients…
TLDR: In this blogpost, we're going to train a neural network that is fully encrypted during training (trained on unencrypted data). The result will be a neural network with two beneficial properties. First, the neural network's intelligence is protected from those who might want to steal it, allowing valuable AIs to be trained in insecure environments without risking theft of their intelligence.…
EDIT: A complete revamp of PyTorch was released today (Jan 18, 2017), making this blogpost a bit obselete. I will update this post with a new Quickstart Guide soon, but for now you should check out their documentation.</a> This Blogpost Will Cover: Part 1: PyTorch Installation Part 2: Matrices and Linear Algebra in PyTorch Part 3: Building a Feedforward Network (starting with a familiar one ) Part…
If you passed high school math and can hack around in Python, I want to teach you Deep Learning . Edit: 50% Coupon Code: "mltrask" (expires August 26) I've decided to write a Deep Learning book in the same style as my blog, teaching Deep Learning from an intuitive perspective, all in Python, using only numpy. I wanted to make the lowest possible barrier to entry to learn Deep Learning. Here's what…
Summary: I learn best with toy code that I can play with. This tutorial teaches DeepMind's Neural Stack machine via a very simple toy example, a short python implementation. I will also explain my thought process along the way for reading and implementing research papers from scratch, which I hope you will find useful. I typically tweet out new blogposts when they're complete at @iamtrask . Feel…
Summary: I learn best with toy code that I can play with. This tutorial teaches Recurrent Neural Networks via a very simple toy example, a short python implementation. Chinese Translation Korean Translation I'll tweet out (Part 2: LSTM) when it's complete at @iamtrask . Feel free to follow if you'd be interested in reading it and thanks for all the feedback! Just Give Me The Code: import copy,…
Summary: Dropout is a vital feature in almost every state-of-the-art neural network implementation. This tutorial teaches how to install Dropout into a neural network in only a few lines of Python code. Those who walk through this tutorial will finish with a working Dropout implementation and will be empowered with the intuitions to install it and tune it in any neural network they encounter.…