Hoeffding's lemma

In probability theory, Hoeffding's lemma is an inequality that bounds the moment-generating function of any bounded random variable. It is named after the Finnish–American mathematical statistician Wassily Hoeffding.

The proof of Hoeffding's lemma uses Taylor's theorem and Jensen's inequality. Hoeffding's lemma is itself used in the proof of McDiarmid's inequality.

Statement of the lemma

Let X be any real-valued random variable with expected value $\mathbb {E} (X)=0$ and such that $a\leq X\leq b$ almost surely. Then, for all $\lambda \in \mathbb {R}$ ,

\mathbb {E} \left[e^{\lambda X}\right]\leq \exp \left({\frac {\lambda ^{2}(b-a)^{2}}{8}}\right).

Note that by the assumption that the random variable $X$ has zero expectation, the $a$ and $b$ in the lemma must satisfy $a\leq 0$ and $0\leq b$ .

Proof of the lemma

(The alternate proof is essentially the same proof with more details fleshed out).

Since $e^{\lambda x}$ is a convex function of $x$ , we have

e^{\lambda x}\leq {\frac {b-x}{b-a}}e^{\lambda a}+{\frac {x-a}{b-a}}e^{\lambda b}\qquad \forall a\leq x\leq b

So, $\mathbb {E} \left[e^{\lambda X}\right]\leq {\frac {b-\mathbb {E} (X)}{b-a}}e^{\lambda a}+{\frac {\mathbb {E} (X)-a}{b-a}}e^{\lambda b}.$

Let $h=\lambda (b-a)$ , $p={\frac {-a}{b-a}}$ and $L(h)=-hp+\ln(1-p+pe^{h})$

Then, ${\frac {b-\mathbb {E} (X)}{b-a}}e^{\lambda a}+{\frac {\mathbb {E} (X)-a}{b-a}}e^{\lambda b}=e^{L(h)}$ since $\mathbb {E} (X)=0$

Taking derivative of $L(h)$ ,

L(0)=L^{'}(0)=0{\text{ and }}L^{''}(h)\leq {\frac {1}{4}}

for all h.

By Taylor's expansion,

$L(h)\leq {\frac {1}{8}}h^{2}={\frac {1}{8}}\lambda ^{2}(b-a)^{2}$

Hence, $\mathbb {E} \left[e^{\lambda X}\right]\leq e^{{\frac {1}{8}}\lambda ^{2}(b-a)^{2}}$

(The "alternative proof" below is the same proof with more explanation.)

Alternative proof

First note that if one of $a$ or $b$ is zero, then $\textstyle {\mathbb {P}}\left(X=0\right)=1$ and the inequality follows. If both are nonzero, then $a$ must be negative and $b$ must be positive.