Cryptography on jeffcarp

Doing Post-Quantum Cryptography in JAX

Sun, 05 Jan 2025 00:00:00 +0000

In 2018 while studying ML and cryptography around the same time, I realized that many cryptographic algorithms can be expressed as computation graphs, the same ones supported by major ML frameworks, which led to a completely frivolous attempt to implement cryptographic algorithms in TensorFlow, just to see if it would work.

The world has changed a lot since 2018. JAX is growing in popularity as an ML framework, and in the cryptography space, post-quantum cryptography has gone from a mostly theoretical threat to a slightly more real one. So continuing the motivation of the original post, I decided to look into implementing a post-quantum cryptography algorithm in JAX.

Doing Cryptography in TensorFlow

Sat, 23 Jun 2018 13:00:14 -0700

On the left: the Feistel Network from the DES cipher, implemented below. On the right: a deep neural network.

TensorFlow is a popular machine learning framework. If you look under the hood, TensorFlow is a general platform for doing computation over tensors in the structure of a graph.

Understanding the Security of Cryptographic Hash Functions

Mon, 16 Apr 2018 21:33:26 -0700

Hash functions are an extremely versatile tool that you can find nearly everywhere in software engineering.

Bitcoin mining is done by repeatedly computing SHA256 hashes.
git stores files internally using the SHA-1 hash of the content as the filename (content-based addressing).
When you download an app, your phone verifies that the app you’re about to run is the app you intended to download by computing a checksum (a hash of the entire app) and comparing it to one provided by the app store. If the values are different, that means someone might have changed the contents of the app.

This article intends to give you a deeper understanding of cryptographic hash functions, when to use them, and how to think about their security properties.

Thoughts on Stanford's CS255 Intro to Cryptography

Sat, 31 Mar 2018 12:31:22 -0700

I’ve always been interested in Cryptography – the field is a really compelling marriage of beautiful mathematics with critical real-world applications. Much of the modern internet would not be possible without the clever math behind asymmetric encryption.

This Winter quarter I took Stanford CS255: Introduction to Cryptography. To be honest, it was quite challenging as I was still working full time as a software engineer on Chrome, and I was never that strong in the math department. However, I got a ton out of this course and solidly recommend it to anyone interested. The professor – Dan Boneh – was amazing and made the complex topics approachable. He publishes the textbook for the course for free here. Here are some interesting things I picked up from the course that I want to take with me.