Hoeffding's lemma
In probability theory, Hoeffding's lemma is an inequality that bounds the moment-generating function of any bounded random variable.[1] 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 and such that almost surely. Then, for all ,
Note that because of the assumption that the random variable
has zero expectation, the
and
in the lemma must satisfy
.
A brief proof of the lemma
Since
is a convex function of
, we have
So,
Let , and
Then, since
Taking derivative of ,
- for all h.
By Taylor's expansion,
Hence,
(The proof below is the same proof with more explanation.)
More detailed proof
First note that if one of or is zero, then and the inequality follows. If both are nonzero, then must be negative and must be positive.
Next, recall that is a convex function on the real line:
Applying to both sides of the above inequality gives us:
Let and define:
is well defined on , to see this we calculate:
The definition of implies
By Taylor's theorem, for every real there exists a between and such that
Note that:
Therefore,
This implies
See also
Notes
- ↑ Pascal Massart (26 April 2007). Concentration Inequalities and Model Selection: Ecole d'Eté de Probabilités de Saint-Flour XXXIII - 2003. Springer. p. 21. ISBN 978-3-540-48503-2.