Jeremy Kun

startups The software world is always atwitter with predictions on the next big piece of technology. And a lot of chatter focuses on what venture capitalists express interest in. As an investor, how do you pick a good company to invest in? Do you notice quirky names like “Kaggle” and “Meebo,” require deep technical abilities, or value a charismatic sales pitch?

Optimism in the Face of Uncertainty: the UCB1 Algorithm

This primer exists for the background necessary to read our post on RSA encryption, but it also serves as a general primer to number theory. Oh, Numbers, Numbers, Numbers We start with some easy definitions.

Number Theory—A Primer

My four-year-old son has declared 36 to be the best number. His reason: 36 is the only number (he knows of) that is both a square and a staircase number AND an up-and-down-staircase number. “Staircase numbers” are what he calls triangular numbers (numbers that are the sum of the first $n$ integers). This name comes from the blocks he has that can be arranged into a staircase. He also calls them “step squad” numbers thanks to Numberblocks.

My 4-year-old declares 36 the best number

A colleague of mine recently lent a hand implementing a polynomial approximation routine I could port to our compiler, though it wasn’t the method I was expecting. As I had written about previously, I was studying the Remez algorithm and implementing a prototype in Python. Remez approximation involves an iterated loop that alternates between root-finding and linear-system solving, and as such it can be rather brittle and difficult.

Anyone know of an approachable proof of the Caratheodory-Fejer theorem?

I’ll be at the Joint Mathematics Meeting in Seattle (starting tomorrow). If you see me there, say hi! I will have a very light schedule, plenty of time for coffee chats. I’ll be attending many of the crypto sessions for the homomorphic encryption talks. And on Thursday at 3PM, I’ll be at the Code4Math booth in the exhibition hall.

I'll be at the JMM

Back in 2020, when I worked in the supply chain side of Google, I had a fun and impactful side project related to human-level explanations of linear programs. A linear program is a mathematical model that defines some number of variables, linear constraints, and a linear objective function. When some variables are forced to be integer (ILPs), you can solve a lot of useful problems like scheduling, routing, and packing.

Explainable Linear Programs

The Hyperfixed Podcast had a lovely episode recently about tape measures. It started from “why does my tape measure seem to always be off a little bit” and went all the way to the inherent limitations of physical measurement at small scales. In there is an awesome quote by Adam Savage, “I had always had faith in the sanctity and solidity of numbers… and when I got into this… I realized there’s no such thing as a measurement.

Epiphanies from Tape Measures

In this living document, I will document reactions to uses of homomorphic encryption by members of the public. By “member of the public,” I mean people who may be technical, but are not directly involved in the development or deployment of homomorphic encryption systems. This includes journalists, bloggers, aggregator comment threads, and social media posts.

Fully Homomorphic Encryption and the Public

In my little corner of the FHE world, things have been steadily heating up. For those who don’t know, my main work project right now is HEIR (Homomorphic Encryption Intermediate Representation), a compiler toolchain for fully homomorphic encryption (FHE). For an extended introduction see this talk from October 2023. The primary focus of HEIR is to compile to FHE hardware accelerators. And boy there are a lot of them.

The last few months in HEIR

Kristin Lauter and her colleagues at Facebook research recently announced a project to benchmark attacks against LWE. The announcement was on the post-quanum crypto mailing list. They state: “Our approach is motivated by the need to study more carefully the effect on security of using small secrets and small error in standardized LWE settings like Kyber and Homomorphic Encryption.

LWE Attack Benchmarking Project

Editor’s note: This essay was originally published in 2019. I have made minor edits in this republishing. There was a MathOverflow thread about mathematically interesting games for 5–6 year olds.

Attention spans for math and stories

This article will explain how the blog is organized at a technical level, and show how I implemented various IndieWeb features.

How This Blog Does IndieWeb

In my recent overview of homomorphic encryption, I underemphasized the importance of data layout when working with arithmetic (SIMD-style) homomorphic encryption schemes. In the FHE world, the name given to data layout strategies is called “packing,” because it revolves around putting multiple plaintext data into RLWE ciphertexts in carefully-chosen ways that mesh well with the operations you’d like to perform.

Packing Matrix-Vector Multiplication in Fully Homomorphic Encryption

This is a story about a failure to apply dynamic programming to a woodworking project. I’ve been building a shed in my backyard, and for one section I decided to build the floor by laying 2x4 planks side by side. I didn’t feel the need to join them with tongue-and-groove, but I did notice that using 2x4s alone wouldn’t fit the width they were supposed to fill.

Dynamic Programming Fail

Table of Contents In this article I’ll show how to use PDLL, a tool for defining MLIR patterns, which itself is built with MLIR. PDLL is intended to be a replacement for defining patterns in tablegen, though there are few public examples of its use.

MLIR — Defining Patterns with PDLL

I’ve been upstreaming a bit of my compiler work to the MLIR project. Yesterday, I merged in a tutorial on mlir-opt, the main debugging tool for running passes on MLIR code. This is roughly the upstreamable parts of my first MLIR tutorial entry, MLIR — Running and Testing a Lowering. Mehdi Amini also provided a lot of useful information during review that taught me some stuff I didn’t know about the tool.

Polynomial dialect and mlir-opt tutorial upstreamed

In this living document, I will list all production systems I’m aware of that use fully homomorphic encryption (FHE). For background on FHE, see my overview of the field. If you have any information about production FHE systems not in this list, or corrections to information in this list, please send me an email with sufficient detail allow the claim to be publicly verified.

Fully Homomorphic Encryption in Production Systems

Ben Recht, a computer science professor at UC Berkeley, recently wrapped up a 3-month series of blog posts on Paul Meehl’s “Philosophical Psychology.” Recht has a table of contents for his blog series. It loosely tracks a set of lectures that Meehl gave in 1989 at the University of Minnesota. In it, he surveys of the philosophy of science, lays out a framework for scientific debate, and critiques scientific practice.

Ben Recht on Meehl's Philosophical Psychology

I added Twitter syndication, and because I have nothing to test it with I’ll share some random life updates. My daughter was born recently, which means I’m on paternity leave for a few months. Hopefully in the liminal hours of sleep training, I’ll have some time to work on my book. Or at least catch up on reading. I finally published my grandmother’s autobiography.

Random life updates

A quick note: you can use C++11 templates to detect struct fields by name and type, and statically branch on them. I first heard of this solution from breeze1990. Say I want to detect if a struct has a field size of type int. Create two template instantiations of the same name, here HasStaticSize that defaults to false.

Detecting field names with C++ metaprogramming

Steven Clontz informed me of an effort he’s involved in called code4math. It’s described as a professional organization for the advancement of mathematical research through building non-research software infrastructure. By that he means, for example, writing software packages like Macaulay2 or databases of mathematical objects that other researchers can use to do their research.

Math databases

POSSE stands for Publish (on your) Own Site, Syndicate Elsewhere. I first heard about it from Cory Doctorow. I’m experimenting with automation to convert posts tagged shortform into Mastodon threads (I’m mathstodon.xyz/@j2kun). I’m using Hugo as a static site generator, with the source a (private) GitHub repository, and Netlify for deployments.

Experiments with POSSE

It’s April Cools! Last year I wrote about friendship bracelets and the year before about cocktails. This year it’s parenting. Parenting articles are a dime a dozen and always bury the lede behind a long story. I’ll skip that. How to think about your child and your role as a parent These are framing devices. Concrete things to do to work toward these are in the next section.

Unusual Tips for Parenting Toddlers

I’ve been learning recently about how to approximate functions by low-degree polynomials. This is useful in fully homomorphic encryption (FHE) in the context of “arithmetic FHE” (see my FHE overview article), where the computational model makes low-degree polynomials cheap to evaluate and non-polynomial functions expensive or impossible. In browsing the state of the art I came across two interesting things.

Remez and function approximations

About two years ago, I switched teams at Google to focus on fully homomorphic encryption (abbreviated FHE, or sometimes HE). Since then I’ve got to work on a lot of interesting projects, learning along the way about post-quantum cryptography, compiler design, and the ins and outs of fully homomorphic encryption.

A High-Level Technical Overview of Fully Homomorphic Encryption

Can you find a set of cards among these six, such that the socks on the chosen cards can be grouped into matching pairs? (Duplicate pairs of the same sock are OK) Spoilers: If the cards are indexed as 1 2 3 4 5 6 Then the following three subsets work: $\{ 1, 2, 4, 5, 6 \}$, $\{ 2, 3, 6 \}$, and $\{ 1, 3, 4, 5 \}$.

Socks, a matching game based on an additive combinatorics problem

There’s a family of tabletop games that are based directly on a nontrivial mathematics problem. As a casual and fun way to inaugurate my new blog (migrated from Wordpress to Hugo, after my work on getting better LaTeX mathmode support in Hugo), I thought I’d write a short listicle about them, so that I have a place to add more as I find them, as well as give the shortest canonical description of the associated math problem.

Tabletop Games Based on Math Problems

Table of Contents In the last article we lowered our custom poly dialect to standard MLIR dialects. In this article we’ll continue lowering it to LLVM IR, exporting it out of MLIR to LLVM, and then compiling to x86 machine code. The code for this article is in this pull request, and as usual the commits are organized to be read in order. Defining a Pipeline The first step in lowering to machine code is to lower to an “exit dialect.

MLIR — Lowering through LLVM

Table of Contents In previous articles we defined a dialect, and wrote various passes to optimize and canonicalize a program using that dialect. However, one of the main tenets of MLIR is “incremental lowering,” the idea that there are lots of levels of IR granularity, and you incrementally lower different parts of the IR, only discarding information when it’s no longer useful for optimizations.

MLIR — Dialect Conversion

Table of Contents In a previous article we defined folding functions, and used them to enable some canonicalization and the sccp constant propagation pass for the poly dialect. This time we’ll see how to add more general canonicalization patterns. The code for this article is in this pull request, and as usual the commits are organized to be read in order. Why is Canonicalization Needed?

MLIR — Canonicalizers and Declarative Rewrite Patterns

In cryptography, we need a distinction between a cleartext and a plaintext. A cleartext is a message in its natural form. A plaintext is a cleartext that is represented in a specific way to prepare it for encryption in a specific scheme. The process of taking a cleartext and turning it into a plaintext is called encoding, and the reverse is called decoding. In homomorphic encryption, the distinction matters.

Encoding Schemes in FHE

Table of Contents Last time we saw how to use pre-defined MLIR traits to enable upstream MLIR passes like loop-invariant-code-motion to apply to poly programs. We left out -sccp (sparse conditional constant propagation), and so this time we’ll add what is needed to make that pass work. It requires the concept of folding. The code for this article is in this pull request, and as usual the commits are organized to be read in order.

MLIR — Folders and Constant Propagation

Table of Contents In the last article in the series, we migrated the passes we had written to use the tablegen code generation framework. That was a preface to using tablegen to define dialects.

MLIR — Defining a New Dialect

Table of Contents As we announced recently, my team at Google has started a new effort to build production-worthy engineering tools for Fully Homomorphic Encryption (FHE). One focal point of this, and one which I’ll be focusing on as long as Google is willing to pay me to do so, is building out a compiler toolchain for FHE in the MLIR framework (Multi-Level Intermediate Representation). The project is called Homomorphic Encryption Intermediate

MLIR — Getting Started

Before I discovered math, I was a first year undergrad computer science student taking Electrical Engineering 101. The first topic I learned was what bits and boolean gates are, and the second was the two’s complement representation of a negative n-bit integer. At the time two’s complement seemed to me like a bizarre quirk of computer programming, with minutiae you just had to memorize.

Two's Complement and Group Theory

In this article I’ll derive a trick used in FHE called sample extraction. In brief, it allows one to partially convert a ciphertext in the Ring Learning With Errors (RLWE) scheme to the Learning With Errors (LWE) scheme. Here are some other articles I’ve written about other FHE building blocks, though they are not prerequisites for this article.

Sample Extraction from RLWE to LWE

This article was written by my colleague, Cathie Yun. Cathie is an applied cryptographer and security engineer, currently working with me to make fully homomorphic encryption a reality at Google. She’s also done a lot of cool stuff with zero knowledge proofs. In previous articles, we’ve discussed techniques used in Fully Homomorphic Encryption (FHE) schemes.

Estimating the Security of Ring Learning with Errors (RLWE)

Problem: Compute the product of two polynomials efficiently. Solution: import numpy from numpy.fft import fft, ifft def poly_mul(p1, p2): """Multiply two polynomials. p1 and p2 are arrays of coefficients in degree-increasing order. """ deg1 = p1.shape[0] - 1 deg2 = p1.shape[0] - 1 # Would be 2*(deg1 + deg2) + 1, but the next-power-of-2 handles the +1 total_num_pts = 2 * (deg1 + deg2) next_power_of_2 = 1 &lt;&lt; (total_num_pts - 1).

Polynomial Multiplication Using the FFT

In this article I’ll cover three techniques to compute special types of polynomial products that show up in lattice cryptography and fully homomorphic encryption. Namely, the negacyclic polynomial product, which is the product of two polynomials in the quotient ring $\mathbb{Z}[x] / (x^N + 1)$. As a precursor to the negacyclic product, we’ll cover the simpler cyclic product. All of the Python code written for this article is on GitHub.

Negacyclic Polynomial Multiplication

Welcome to the 209th Carnival of Mathematics!

Carnival of Mathematics #209

Last time we covered an operation in the LWE encryption scheme called modulus switching, which allows one to switch from one modulus to another, at the cost of introducing a small amount of extra noise, roughly $\sqrt{n}$, where $n$ is the dimension of the LWE ciphertext.

Key Switching in LWE

It’s April Cools! We’re taking back April Fools. When I was younger I had a strange relationship with alcohol, not because of any sort of trauma, but because I found it decidedly boring and disgusting to the taste. I didn’t drink in high school, didn’t enjoy parties in college, and didn’t care for tailgating or other sports-based events where drinking was common.

Cocktails

tl;dr: I’m writing a new book, sign up for the announcements mailing list. I’ve written exactly zero new technical blog posts this year because I’ve been spending all my writing efforts on my next book, Practical Math for Programmers (PMFP, subtitle: A Tour of Mathematics in Production Software). I’ve written a little bit about it in my newsletter, Halfspace. There I rant, critique, brainstorm, and wax poetic about math and software.

My next book will be "Practical Math for Programmers"

Lately I’ve been studying Fully Homomorphic Encryption, which is the miraculous ability to perform arbitrary computations on encrypted data without learning any information about the underlying message.

The Gadget Decomposition in FHE

Welcome to the 197th Carnival of Mathematics! 197 is an unseemly number, as you can tell by the Wikipedia page which currently says that it has “indiscriminate, excessive, or irrelevant examples.” How deviant. It’s also a Repfigit, which means if you start a fibonacci-type sequence with the digits 1, 9, 7, and then continue with $ a_n = a_{i-3} + a_{i-2} + a_{i-1}$, then 197 shows up in the sequence.

Carnival of Mathematics #197

Recently I’ve been helping out with a linear algebra course organized by Tai-Danae Bradley and Jack Hidary, and one of the questions that came up a few times was, “why should programmers care about the concept of a linear combination?” For those who don’t know, given vectors $ v_1, \dots, v_n$, a linear combination of the vectors is a choice of some coefficients $ a_i$ with which to weight the vectors in a sum $ v = \sum_{i=1}^n a_i v_i$.

Regression and Linear Combinations

We’re ironically searching for counterexamples to the Riemann Hypothesis. Setting up Pytest Adding a Database Search Strategies Unbounded integers Deploying with Docker Performance Profiling Last time we made the audacious choice to remove primary keys from the RiemannDivisorSums table for performance reasons.

Searching for RH Counterexamples — Scaling Up

We’re ironically searching for counterexamples to the Riemann Hypothesis. Setting up Pytest Adding a Database Search Strategies Unbounded integers Deploying with Docker In the last article we ran into some performance issues with our deployed docker application.

Searching for RH Counterexamples — Performance Profiling

We’re ironically searching for counterexamples to the Riemann Hypothesis. Setting up Pytest Adding a Database Search strategies In the last article, we improved our naive search from “try all positive integers” to enumerate a subset of integers (superabundant numbers), which RH counterexamples are guaranteed to be among. These numbers grow large, fast, and we quickly reached the limit of what 64 bit integers can store.

Searching for RH Counterexamples — Unbounded Integers

We’re glibly searching for counterexamples to the Riemann Hypothesis, to trick you into learning about software engineering principles. In the first two articles we configured a testing framework and showed how to hide implementation choices behind an interface. Next, we’ll improve the algorithm’s core routine. As before, I’ll link to specific git commits in the final code repository to show how the project evolves.

Searching for RH Counterexamples — Search Strategies

Some mathy-programmy people tell me they want to test their code, but struggle to get set up with a testing framework. I suspect it’s due to a mix of: There are too many choices with a blank slate. Making slightly wrong choices early on causes things to fail in unexpected ways. I suspect the same concerns apply to general project organization and architecture.

Searching for RH Counterexamples — Setting up Pytest

In my book, A Programmer’s Introduction to Mathematics, I describe the Taylor Series as a “hammer for every nail.” I learned about another nail in the design of modern smartphone accelerometers from “Eight Amazing Engineering Stories” by Hammack, Ryan, and Ziech, which I’ll share here. These accelerometers are designed using a system involving three plates, which correspond to two capacitors.

Taylor Series and Accelerometers

In my book I discuss the importance of context in reading and writing mathematics. An early step in becoming comfortable with math is deciphering the syntax of mathematical expressions. Another is in connecting the symbols to their semantic meanings. Embedded in these is the subproblem of knowing what to call the commonly used symbols. The more abstract you go, the more exotic the symbols tend to get.

Contextual Symbols in Math

This essay is a slightly modified version of the closing chapter of A Programmer’s Introduction to Mathematics. We are no longer constrained by pencil and paper. The symbolic shuffle should no longer be taken for granted as the fundamental mechanism for understanding quantity and change. Math needs a new interface. –Bret Victor, “Kill Math” Math is a human activity.

Musings on A New Interface for Mathematics

Recently my employer (Google) forced me to switch to Mercurial instead of my usual version control system, git. The process of switching sparked a few discussions between me and my colleagues about the value of various version control systems. A question like “what benefit does git provide over Mercurial” yielded no clear answers, suggesting many developers don’t know. An informal Twitter survey didn’t refute this claim.

The Communicative Value of Using Git Well

Previous posts in this series: Silent Duels and an Old Paper of Restrepo Silent Duels—Parsing the Construction Last time we waded into Restrepo’s silent duel paper. You can see the original and my re-typeset version on Github along with all of the code in this series. We digested Section 2 and a bit further, plotting some simplified examples of the symmetric version of the game. I admit, this paper is not an easy read.

Silent Duels—Constructing the Solution part 1

Our hero, a mathematician, is writing notes in LaTeX and needs to convert it to a format that her blog platform accepts. She’s used to using dollar sign delimiters for math mode, but her blog requires  and \[ \]. Find-and-replace fails because it doesn’t know about which dollar sign is the start and which is the end. She knows there’s some computer stuff out there that could help, but she doesn’t have the damn time to sort through it all.

A Working Mathematician's Guide to Parsing

Last time we discussed the setup for the silent duel problem: two players taking actions in $ [0,1]$, player 1 gets $ n$ chances to act, player 2 gets $ m$, and each knows their probability of success when they act. The solution is in a paper of Rodrigo Restrepo from the 1950s. In this post I’ll start detailing how I study this paper, and talk through my thought process for approaching a bag of theorems and proofs.

Silent Duels—Parsing the Construction

For the last four years I’ve been working on a book for programmers who want to learn mathematics. It’s finally done, and you can buy it today. The website for the book is pimbook.org, which has purchase links—paperback and ebook—and a preview of the first pages. You can see more snippets later in the book on the Amazon listing’s “Look Inside” feature. If you’re a programmer who wants to learn math, this book is written specifically for you!

A Programmer's Introduction to Mathematics

Over at Math3ma, Tai-Danae Bradley shared the following puzzle, which she also featured in a fantastic (spoiler-free) YouTube video. If you’re seeing this for the first time, watch the video first. Consider a square in the xy-plane, and let A (an “assassin”) and T (a “target”) be two arbitrary-but-fixed points within the square.

Visualizing an Assassin Puzzle

Problem: Compute distance between points with uncertain locations (given by samples, or differing observations, or clusters). For example, if I have the following three “points” in the plane, as indicated by their colors, which is closer, blue to green, or blue to red? It’s not obvious, and there are multiple factors at work: the red points have fewer samples, but we can be more certain about the position;

Earthmover Distance

Every now and then I hear some ridiculous things about the equals symbol. Some large subset of programmers—perhaps related to functional programmers, perhaps not—seem to think that = should only and ever mean “equality in the mathematical sense.” The argument usually goes, Functional programming gives us back that inalienable right to analyze things by using mathematics.

For mathematicians, = does not mean equality

When NP-hardness pops up on the internet, say because some silly blogger wants to write about video games, it’s often tempting to conclude that the problem being proved NP-hard is actually very hard! “Scientists proved Super Mario is NP-hard? I always knew there was a reason I wasn’t very good at it!” Sorry, these two are unrelated. NP-hardness means hard in a narrow sense this post should hopefully make clear.

NP-hard does not mean hard

The hypothesis and the setup This blog post has an interactive demo (mostly used toward the end of the post). The source for this demo is available in a Github repository. Last time we saw how the inner product of two vectors gives rise to a decision rule: if $ w$ is the normal to a line (or hyperplane) $ L$, the sign of the inner product $ \langle x, w \rangle$ tells you whether $ x$ is on the same side of $ L$ as $ w$.

Formulating the Support Vector Machine Optimization Problem

Binary search is one of the most basic algorithms I know. Given a sorted list of comparable items and a target item being sought, binary search looks at the middle of the list, and compares it to the target. If the target is larger, we repeat on the smaller half of the list, and vice versa. With each comparison the binary search algorithm cuts the search space in half.

Binary Search on Graphs

Previously in this series: Linear programming and healthy diets — Part 1 Linear programing and the simplex algorithm Foods of the Father My dad’s an interesting guy. Every so often he picks up a health trend and/or weight loss goal that would make many people’s jaw drop.

Linear Programming and Healthy Diets — Part 2

Last week I was in Boston for the Geometry of Redistricting workshop. It was an optimistic gathering of over 500 mathematicians, computer scientists, lawyers, policy makers, teachers, and interested people of all stripes. There was a ton of information in the talks and subsequent discussions. I’ll try to distill the main ideas and avenues for research as best I can.

Notes on Math and Gerrymandering

Problem: Determine if two polynomial expressions represent the same function. Specifically, if $ p(x_1, x_2, \dots, x_n)$ and $ q(x_1, x_2, \dots, x_n)$ are a polynomial with inputs, outputs and coefficients in a field $ F$, where $ |F|$ is sufficiently large, then the problem is to determine if $ p(\mathbf{x}) = q(\mathbf{x})$ for every $ x \in F$, in time polynomial in the number of bits required to write down $ p$ and $ q$.

Testing Polynomial Equality

Problem: You have a catalog of items with discrete ratings (thumbs up/thumbs down, or 5-star ratings, etc.), and you want to display them in the “right” order.

Bayesian Ranking for Rated Items

papad Hard to believe Sanjeev Arora and his coauthors consider it “a basic tool [that should be] taught to all algorithms students together with divide-and-conquer, dynamic programming, and random sampling.” Christos Papadimitriou calls it “so hard to believe that it has been discovered five times and forgotten.” It has formed the basis of algorithms in machine learning, optimization, game theory, economics, biology, and more.

The Reasonable Effectiveness of the Multiplicative Weights Update Algorithm

So far in this series we’ve seen a lot of motivation and defined basic ideas of what a quantum circuit is. But on rereading my posts, I think we would all benefit from some concreteness. “Local” operations So by now we’ve understood that quantum circuits consist of a sequence of gates $ A_1, \dots, A_k$, where each $ A_i$ is an 8-by-8 matrix that operates “locally” on some choice of three (or fewer) qubits.

Concrete Examples of Quantum Gates

This is a guest post by my friend and colleague Samantha Davies. Samantha is a math Ph.D student at the University of Washington, and a newly minted math blogger. Go check out her blog, With High Probability. If I said “let’s talk about temperature and voltage”, you might be interested, but few would react the same if instead I suggested an umbrella term: harmonic functions.

Voltage, Temperature, and Harmonic Functions

The next Monday, when the fathers were all back at work, we kids were playing in a field. One kid says to me, “See that bird? What kind of bird is that?” I said, “I haven’t the slightest idea what kind of a bird it is.” He says, “It’s a brown-throated thrush. Your father doesn’t teach you anything!” But it was the opposite. He had already taught me: “See that bird?

Zero-Knowledge: Definitions and Theory

Welcome to the 233rd Carnival of Mathematics! Who can forget 233, the 6th Fibonacci prime? Hey, not all numbers are interesting. Don’t ask me about the smallest positive uninteresting number. You can’t make it interesting with your feeble mind tricks! Anyway, on to the fun. Provers and Shakers The big discovery this month was a new largest known prime number, $2^{136279841} - 1$, as reported by the Great Internet Mersenne Prime Search.

Carnival of Mathematics #233

In this post we’ll get a strong taste for zero knowledge proofs by exploring the graph isomorphism problem in detail. In the next post, we’ll see how this relates to cryptography and the bigger picture. The goal of this post is to get a strong understanding of the terms “prover,” “verifier,” and “simulator,” and “zero knowledge” in the context of a specific zero-knowledge proof.

Zero Knowledge Proofs — A Primer

The singular value decomposition (SVD) of a matrix is a fundamental tool in computer science, data analysis, and statistics. It’s used for all kinds of applications from regression to prediction, to finding approximate solutions to optimization problems. In this series of two posts we’ll motivate, define, compute, and use the singular value decomposition to analyze some data.

Singular Value Decomposition Part 1: Perspectives on Linear Algebra

Variations on a theme Back in 2014 I wrote a post called How to Conquer Tensorphobia that should end up on Math $ \cap$ Programming’s “greatest hits” album. One aspect of tensors I neglected to discuss was the connection between the modern views of tensors and the practical views of linear algebra. I feel I need to write this because every year or two I forget why it makes sense.

Tensorphobia and the Outer Product

Problem: Estimate the number of distinct items in a data stream that is too large to fit in memory.

Hashing to Estimate the Size of a Stream

Math and computer science are full of inequalities, but there is one that shows up more often in my work than any other. Of course, I’m talking about $$\displaystyle 1+x \leq e^{x}$$ This is The Inequality. I’ve been told on many occasions that the entire field of machine learning reduces to The Inequality combined with the Chernoff bound (which is proved using The Inequality). Why does it show up so often in machine learning?

The Inequality

Here’s a bit of folklore I often hear (and retell) that’s somewhere between a joke and deep wisdom: if you’re doing a software interview that involves some algorithms problem that seems hard, your best bet is to use hash tables. More succinctly put: Google loves hash tables. As someone with a passion for math and theoretical CS, it’s kind of silly and reductionist.

Load Balancing and the Power of Hashing

I was recently an invited speaker in a series of STEM talks at Moraine Valley Community College. My talk was called “What can algorithms tell us about life, love, and happiness?” and it’s on Youtube now so you can go watch it. The central theme of the talk was the lens of computation, that algorithms and theoretical computer science can provide new and novel explanations for the natural phenomena we observe in the world.

Serial Dictatorships and House Allocation

There’s a well-understood phenomenon in machine learning called overfitting. The idea is best shown by a graph: overfitting Let me explain. The vertical axis represents the error of a hypothesis. The horizontal axis represents the complexity of the hypothesis. The blue curve represents the error of a machine learning algorithm’s output on its training data, and the red curve represents the generalization of that hypothesis to the real world.

The Boosting Margin, or Why Boosting Doesn't Overfit

In 2014 the White House commissioned a 90-day study that culminated in a report (pdf) on the state of “big data” and related technologies. The authors give many recommendations, including this central warning. Warning: algorithms can facilitate illegal discrimination! Here’s a not-so-imaginary example of the problem. A bank wants people to take loans with high interest rates, and it also serves ads for these loans.

What does it mean for an algorithm to be fair?

In this post we’ll implement Reed-Solomon error-correcting codes and use them to play with codes. In our last post we defined Reed-Solomon codes rigorously, but in this post we’ll focus on intuition and code. As usual the code and data used in this post is available on this blog’s Github page.

The Welch-Berlekamp Algorithm for Correcting Errors in Data

When addressing the question of what it means for an algorithm to learn, one can imagine many different models, and there are quite a few. This invariably raises the question of which models are “the same” and which are “different,” along with a precise description of how we’re comparing models.

Weak Learning, Boosting, and the AdaBoost algorithm

A while back Peter Norvig posted a wonderful pair of articles about regex golf. The idea behind regex golf is to come up with the shortest possible regular expression that matches one given list of strings, but not the other. “Regex Golf,” by Randall Munroe. In the first article, Norvig runs a basic algorithm to recreate and improve the results from the comic, and in the second he beefs it up with some improved search heuristics.

The Many Faces of Set Cover

Or how to detect and correct errors Last time we made a quick tour through the main theorems of Claude Shannon, which essentially solved the following two problems about communicating over a digital channel. What is the best encoding for information when you are guaranteed that your communication channel is error free? Are there any encoding schemes that can recover from random noise introduced during transmission?

Hamming's Code

Last time we defined the Hamming code. We also saw that it meets the Hamming bound, which is a measure of how densely a code can be packed inside an ambient space and still maintain a given distance. This time we’ll define the Reed-Solomon code which optimizes a different bound called the Singleton bound, and then generalize them to a larger class of codes called Reed-Muller codes.

The Codes of Solomon, Reed, and Muller

There are two basic problems in information theory that are very easy to explain. Two people, Alice and Bob, want to communicate over a digital channel over some long period of time, and they know the probability that certain messages will be sent ahead of time. For example, English language sentences are more likely than gibberish, and “Hi” is much more likely than “asphyxiation.” The problems are: Say communication is very expensive.

A Proofless Introduction to Information Theory

Last time we left off with a tantalizing conjecture: a random graph with edge probability $ p = 5/n$ is almost surely a connected graph. We arrived at that conjecture from some ad-hoc data analysis, so let’s go back and treat it with some more rigorous mathematical techniques.

The Giant Component and Explosive Percolation

Greedy algorithms are by far one of the easiest and most well-understood algorithmic techniques. There is a wealth of variations, but at its core the greedy algorithm optimizes something using the natural rule, “pick what looks best” at any step. So a greedy routing algorithm would say to a routing problem: “You want to visit all these locations with minimum travel time? Let’s start by going to the closest one.

When Greedy Algorithms are Perfect: the Matroid

Last time we left off with the tantalizing question: how do you do a quantum “AND” operation on two qubits? In this post we’ll see why the tensor product is the natural mathematical way to represent the joint state of multiple qubits. Then we’ll define some basic quantum gates, and present the definition of a quantum circuit.

Multiple Qubits and the Quantum Circuit

The best place to start our journey through quantum computing is to recall how classical computing works and try to extend it. Since our final quantum computing model will be a circuit model, we should informally discuss circuits first. A circuit has three parts: the “inputs,” which are bits (either zero or one); the “gates,” which represent the lowest-level computations we perform on bits;

The Quantum Bit

Quantum mechanics is one of the leading scientific theories describing the rules that govern the universe. It’s discovery and formulation was one of the most important revolutions in the history of mankind, contributing in no small part to the invention of the transistor and the laser. Here at Math ∩ Programming we don’t put too much emphasis on physics or engineering, so it might seem curious to study quantum physics.

A Motivation for Quantum Computing

Last time we saw the Diffie-Hellman key exchange protocol, and discussed the discrete logarithm problem and the related Diffie-Hellman problem, which form the foundation for the security of most protocols that use elliptic curves. Let’s continue our journey to investigate some more protocols. Just as a reminder, the Python implementations of these protocols are not at all meant for practical use, but for learning purposes.

Sending and Authenticating Messages with Elliptic Curves

Optimization is by far one of the richest ways to apply computer science and mathematics to the real world. Everybody is looking to optimize something: companies want to maximize profits, factories want to maximize efficiency, investors want to minimize risk, the list just goes on and on. The mathematical tools for optimization are also some of the richest mathematical techniques.

Linear Programming and Healthy Diets — Part 1

In a previous post we introduced a learning model called Probably Approximately Correct (PAC). We saw an example of a concept class that was easy to learn: intervals on the real line (and more generally, if you did the exercise, axis-aligned rectangles in a fixed dimension). One of the primary goals of studying models of learning is to figure out what is learnable and what is not learnable in the various models.

A problem that is not (properly) PAC-learnable

Here is a fun puzzle. Suppose we have a group of 10 men and 10 women, and each of the men has sorted the women in order of their preference for marriage (that is, a man prefers to marry a woman earlier in his list over a woman later in the list). Likewise, each of the women has sorted the men in order of marriageability. We might ask if there is any way that we, the omniscient cupids of love, can decide who should marry to make everyone happy.

Stable Marriages and Designing Markets

So far in this series we’ve seen elliptic curves from many perspectives, including the elementary, algebraic, and programmatic ones. We implemented finite field arithmetic and connected it to our elliptic curve code. So we’re in a perfect position to feast on the main course: how do we use elliptic curves to actually do cryptography?

Elliptic Curve Diffie-Hellman

Back when I was first exposed to programming language design, I decided it would be really cool if there were a language that let you define your own number types and then do all your programming within those number types.

Math ∩ Programming

Google's PageRank—Introduction