**Halting Problem**
The problem of determining if a program halts or doesn't halt on
a particular input. This is an incomputable problem.

**Halting Set**
The recursively enumerable set of Gödel numbers that
correspond to programs that halt if given their own Gödel
number as input.

**Hebbian Learning**
A rule that specifies that the strength of a synapse between two
neurons should be proportional to the product of the activations
of the two neurons.
cd

**Hénon Map**
A chaotic system (defined by the two equations *x(t+1) = a
- x(t)^2 + b y(t)* and *y(t+1) = x(t)*) that has a fractal
strange attractor and operates in discrete time.

**Hidden Layer**
In a feedforward or recurrent neural network, a layer
of neurons that is neither the input layer nor the output layer
but is physically between the two.

**Hill-Climbing**
One of the simplest search methods that attempts to find a
local maximum by moving in an uphill direction. It is related
to steepest ascent. Hill-climbing may use gradient
information, or random sampling of nearby points, in order to
estimate the uphill direction.

**Holism**
The idea that ``the whole is greater than the sum of the parts.''
Holism is credible on the basis of emergence alone, since
reductionism and bottom-up descriptions of nature often fail
to predict complex higher-level patterns. See also top-down.

**Hopfield Network**
A type of feedback neural network that is often used as an
associative memory or as a solution to a combinatorial
optimization problem.