perceptrons and neural networks (a generalization of the perceptron) are ways to formulate the hypothesis set
that represents your target function
. There is also a second ingredient in learning, namely how you find a member of your hypothesis set
that fits your data. Particle swarms are a particular way of finding such a
and genetic algorithms also correspond to a related way to find your
when your
is obtained as the combination of simple functions using simple combination rules. In general, the way to find
falls under the topic of optimization, and so particle swarms and genetic algorithms are two ways of finding fits to the data that have "high fitness".
This course does not consider these particular ways to optimize (i.e. fit the data to produce
), but we do consider other optimization techniques based on different ideas.
Quote:
Originally Posted by cjohnson
In addition to perceptrons and neural networks, I have read about particle swarms and genetic algorithms. Will these also be addressed as well, or are they outside the realm of this course?
