A book chapter on privacy (in Norwegian)
Last year I was asked by Norges Tekniske Vitenskapsakademi (The Norwegian Technical Science Academy) to write a chapter on privacy in the upcoming addition ...
2021
2020
Why the \(k\) in \(k\)-anonymity does not reflect any re-identification probability: defeating syntactic privacy
Over the years I’ve repeatedly heard the claim that (k)-anonymity means that the re-identification probability is at most (). In an important sense this is ...
Differentially private \(k\)-anonymity makes limited sense
In response to the COVID-19 pandemic, the Norwegian Institute of Health first commissioned and then deployed a population surveillance / contact tracing cel...
2017
Native advertising and guilt by association
In the last year, I have repeatedly been offered the opportunity to contribute to marketing campaigns produced by the content marketing firm Mediaplanet. Fo...
New site generator
I finally moved my pages from a static page generation system that I cobbled together with git, makefiles, pandoc, and css, to the Jekyll based pages you ar...
A reminder regarding p -values and their use
The use, and usefulness, of (p)-values is controversial. One of the reasons for this is that (p)-values often are misinterpreted. One such misinterpretation...
2015
A quick observation on the provenance of randomness in privacy
Checking out Kifer and Machanavajjhala’s TODS paper on their Pufferfish privacy framework, I read the following (note that (M_1) and (M_2) are randomized qu...
Using convolutional networks for object comparison
Recently Google open-sourced their flow graph based numerical computation library Tensorflow. In their words:
Sorting as a convex optimization problem
Lately, I have been working on problems that require solving convex optimization problems. In the course of this work I have been using cvxpy. The cvxpy hom...
2014
A self-recognizing function
Recently I have been thinking about recursive sets. These are sets that have a computable characteristic function. A standard example of a set that is not r...
2013
What does ε-differential privacy mean?
Differential privacy (Dwork et al. 2006) is an emerging standard for database privacy. As such, differential privacy is concerned with databases describing ...
2011
Folding trees and efficient color coding
Folds are convenient and efficient abstractions of recursive patterns for processing data structures. In 1984 John Hughes wrote “Why functional programming ...
2010
R and closures: a potential problem
R is an environment for exploratory data analysis that is becoming increasingly popular. It is also used for analysis of data in high-impact scientific publ...
Efficient sampling without replacement in Python
Sampling (n) elements without replacement from a collection of (N) elements means that no duplicates are allowed. A strategy for sampling without replacemen...
Memory lane: An old C program – Tic Tac Toe
This very old program of mine, written such that the code fits into a 25 lines 80 columns terminal window, lets you play Tic Tac Toe on a (19 ) grid:
An iterated quine in Python
A Quine is a program that prints its own source code. Writing one is considered a rite of passage for programmers. In his Turing award lecture, “Reflections...
Currying, Composition, and a Monad in Python
To cite Wikipedia: