![]() |
(2) |
For small residuals R, the Huber function reduces to the usual L2 least squares penalty function, and for large R it reduces to the usual robust (noise insensitive) L1 penalty function. Notice the continuity at |R|= h where the Huber function switches from its L2 range to its L1 range. Likewise derivatives are continuous at the junctions |R|=h:
![]() |
(3) |
The derivative of the Huber function
is what we commonly call the clip function .
![]() |
(4) |
Now let us set out to minimize a sum
of Huber functions of all the components of the residual
where the residual
is perturbed by the addition
of a small amount of gradient
and previous step
.The perturbed residual is
where we are given
,
,
, and
we seek to find
and
by setting to zero derivatives of
by
and
.For simplicity we assume that
and
are small
and that we do not need to worry about components jumping between
the L2 and L1 range portions of the Huber function.
Obviously residual component values will often jump between the two ranges,
and because of that, we must iterate the steps I define next:
![]() |
(5) | |
(6) |
![]() |
(7) | |
(8) |
From the economical viewpoint,
whether or not we would
iterate for the values of and
would depend on whether
it was
costly to compute
the new gradient
,that is, whether
and
are costly to apply.
If they are, we would want to make sure we got the
most value from each
we had,
so we would iterate the plane search for
.Otherwise, if it was cheap to compute the next gradient
,we would do so rather than making the best possible use
of the existing gradient (by repeated plane search).
The economical viewpoint may be surpassed by
the need to avoid trouble.
Limited experiences so far show that
instabilities can arise
going from one to the next.
We should be able to control them by
iterating to convergence for each
.Failing in that,
I believe theory says we are assured stable
convergence if we drop back from
conjugate directions to steepest descent.
All these extra precautions
will require more than the straightforward coding below.