Blog posts


Calibration by group, error rate parity, sufficiency, and separation


In the field of algorithmic fairness, it is well known that there are several definitions of fairness that are impossible to reconcile except in (practically irrelevant) corner cases. In this context, I have recently tried to wrap my head around why – intuitively – it is impossible for any classifier to achieve separation and sufficiency at the same time (unless either the classifier is a perfect classifier or there are no base rate differences between groups – we will get to these details in a minute). Since part of my troubles arose from a misunderstanding of what separation and sufficiency actually mean, let us start by revisiting their definitions.

Maximum likelihood, cross-entropy, risk minimization


really, yet another post about about maximum likelihood (ML) estimation? Well – yes; I could not find a source that summarized exactly the things I needed to know, so here it is. What will you find?


Beautiful boxplots in pgfplots


Recently, I wanted to create boxplots from a data file for a paper I was writing using pgfplots. Turns out that’s more difficult than expected, especially since the (otherwise very useful) documentation is a bit meager in this point.