Sample compression schemes were defined by Littlestone and Warmuth (1986) as an abstraction of the structure underlying many learning algorithms. Roughly speaking, a sample compression scheme of size $k$ means that given an arbitrary list of labeled examples, one can retain only $k$ of them in a way that allows to recover the labels of all other examples in the list. They showed that compression implies PAC learnability for binary-labeled classes, and asked whether the other direction holds. We answer their question and show that every concept class $C$ with VC dimension $d$ has a sample compression scheme of size exponential in $d$. The proof uses an approximate minimax phenomenon for binary matrices of low VC dimension, which may be of interest in the context of game theory.

The previous version of this text contained an error; Theorem 2.1 in it is false. This error only affects the statement for multi-labeled classes, and the construction for binary-labeled classes still holds. In the new version of the text, we added a relevant discussion in Section 4.

We prove that proper PAC learnability implies compression. Namely, if a concept $C \subseteq \Sigma^X$ is properly PAC learnable with $d$ samples, then $C$ has a sample compression scheme of size $2^{O(d)}$.
In particular, every boolean concept class with constant VC dimension has a sample compression scheme of constant size. This answers a question of Littlestone and Warmuth (1986). The proof uses an approximate minimax phenomenon for boolean matrices of low VC dimension.